diff --git a/hutool-core/src/main/java/cn/hutool/core/text/dfa/NFA.java b/hutool-core/src/main/java/cn/hutool/core/text/dfa/NFA.java
index 468658a62..3d791679c 100644
--- a/hutool-core/src/main/java/cn/hutool/core/text/dfa/NFA.java
+++ b/hutool-core/src/main/java/cn/hutool/core/text/dfa/NFA.java
@@ -4,142 +4,186 @@ import java.util.*;
 
 /**
  * <p>
- *
  * 基于非确定性有穷自动机（NFA） 实现的多模匹配工具
+ * </p>
  *
  * @author renyp
  */
 public class NFA {
-	private final Node root;
+    /**
+     * AC树的根节点
+     */
+    private final Node root;
+    /**
+     * 标记是否需要构建AC自动机，做树优化
+     */
+    private volatile boolean needBuildAc;
 
-	/**
-	 * 默认构造
-	 */
-	public NFA() {
-		this.root = new Node();
-	}
+    /**
+     * 内置锁，防止并发场景，并行建AC树，造成不可预知结果
+     */
+    private final Object buildAcLock;
 
-	/**
-	 * 构造函数 并 初始化词库
-	 *
-	 * @param words 添加的新词
-	 */
-	public NFA(final String... words) {
-		this();
-		this.insert(words);
-	}
+    /**
+     * 内置锁，防止并行插入，新节点建立后，被挂载到树上前 被篡改
+     */
+    private final Object insertTreeLock;
 
-	/**
-	 * 词库添加新词，初始化查找树
-	 *
-	 * @param word 添加的新词
-	 */
-	public void insert(final String word) {
-		Node p = root;
-		for (final char curr : word.toCharArray()) {
-			if (p.next.get((int) curr) == null) {
-				p.next.put((int) curr, new Node());
-			}
-			p = p.next.get((int) curr);
-		}
-		p.flag = true;
-		p.str = word;
-	}
+    /**
+     * 默认构造
+     */
+    public NFA() {
+        this.root = new Node();
+        this.needBuildAc = true;
+        this.buildAcLock = new Object();
+        this.insertTreeLock = new Object();
+    }
 
-	/**
-	 * 词库批量添加新词，初始化查找树
-	 *
-	 * @param words 添加的新词
-	 */
-	public void insert(final String... words) {
-		for (final String word : words) {
-			this.insert(word);
-		}
-	}
+    /**
+     * 构造函数 并 初始化词库
+     *
+     * @param words 添加的新词
+     */
+    public NFA(String... words) {
+        this();
+        this.insert(words);
+    }
 
-	/**
-	 * 构建基于NFA模型的 AC自动机
-	 */
-	public void buildAc() {
-		final Queue<Node> queue = new LinkedList<>();
-		final Node p = root;
-		for (final Integer key : p.next.keySet()) {
-			p.next.get(key).fail = root;
-			queue.offer(p.next.get(key));
-		}
-		while (!queue.isEmpty()) {
-			final Node curr = queue.poll();
-			for (final Integer key : curr.next.keySet()) {
-				Node fail = curr.fail;
-				// 查找当前节点匹配失败，他对应等效匹配的节点是哪个
-				while (fail != null && fail.next.get(key) == null) {
-					fail = fail.fail;
-				}
-				// 代码到这，有两种可能，fail不为null，说明找到了fail；fail为null，没有找到，那么就把fail指向root节点（当到该节点匹配失败，那么从root节点开始重新匹配）
-				if (fail != null) {
-					fail = fail.next.get(key);
-				} else {
-					fail = root;
-				}
-				curr.next.get(key).fail = fail;
-				queue.offer(curr.next.get(key));
-			}
-		}
-	}
+    /**
+     * 词库添加新词，初始化查找树
+     *
+     * @param word 添加的新词
+     */
+    public void insert(String word) {
+        synchronized (insertTreeLock) {
+            needBuildAc = true;
+            Node p = root;
+            for (char curr : word.toCharArray()) {
+                int ind = curr;
+                if (p.next.get(ind) == null) {
+                    p.next.put(ind, new Node());
+                }
+                p = p.next.get(ind);
+            }
+            p.flag = true;
+            p.str = word;
+        }
+    }
 
-	/**
-	 * @param text 查询的文本（母串）
-	 * @return 关键字列表
-	 */
-	public List<FoundWord> find(final String text) {
-		return this.find(text, true);
-	}
+    /**
+     * 词库批量添加新词，初始化查找树
+     *
+     * @param words 添加的新词
+     */
+    public void insert(String... words) {
+        for (String word : words) {
+            this.insert(word);
+        }
+    }
 
-	/**
-	 * @param text           查找的文本（母串）
-	 * @param isDensityMatch 是否密集匹配
-	 * @return 关键字列表
-	 */
-	public List<FoundWord> find(final String text, final boolean isDensityMatch) {
-		final List<FoundWord> ans = new ArrayList<>();
-		Node p = root, k;
-		for (int i = 0, len = text.length(); i < len; i++) {
-			final int ind = text.charAt(i);
-			// 状态转移(沿着fail指针链接的链表，此处区别于DFA模型)
-			while (p != null && p.next.get(ind) == null) {
-				p = p.fail;
-			}
-			if (p == null) {
-				p = root;
-			} else {
-				p = p.next.get(ind);
-			}
-			// 提取结果(沿着fail指针链接的链表，此处区别于DFA模型)
-			k = p;
-			while (k != null) {
-				if (k.flag) {
-					ans.add(new FoundWord(k.str, k.str, i - k.str.length() + 1, i));
-					if (!isDensityMatch) {
-						p = root;
-						break;
-					}
-				}
-				k = k.fail;
-			}
-		}
-		return ans;
-	}
+    /**
+     * 构建基于NFA模型的 AC自动机
+     */
+    private void buildAc() {
+        Queue<Node> queue = new LinkedList<>();
+        Node p = root;
+        for (Integer key : p.next.keySet()) {
+            p.next.get(key).fail = root;
+            queue.offer(p.next.get(key));
+        }
+        while (!queue.isEmpty()) {
+            Node curr = queue.poll();
+            for (Integer key : curr.next.keySet()) {
+                Node fail = curr.fail;
+                // 查找当前节点匹配失败，他对应等效匹配的节点是哪个
+                while (fail != null && fail.next.get(key) == null) {
+                    fail = fail.fail;
+                }
+                // 代码到这，有两种可能，fail不为null，说明找到了fail；fail为null，没有找到，那么就把fail指向root节点（当到该节点匹配失败，那么从root节点开始重新匹配）
+                if (fail != null) {
+                    fail = fail.next.get(key);
+                } else {
+                    fail = root;
+                }
+                curr.next.get(key).fail = fail;
+                queue.offer(curr.next.get(key));
+            }
+        }
+        needBuildAc = false;
+    }
 
-	private static class Node {
+    /**
+     * @param text 查询的文本（母串）
+     */
+    public List<FoundWord> find(String text) {
+        return this.find(text, true);
+    }
 
-		boolean flag;
-		Node fail;
-		String str;
-		Map<Integer, Node> next;
+    /**
+     * @param text           查找的文本（母串）
+     * @param isDensityMatch 是否密集匹配
+     */
+    public List<FoundWord> find(String text, boolean isDensityMatch) {
+        // double check，防止重复无用的 buildAC
+        if (needBuildAc) {
+            synchronized (buildAcLock) {
+                if (needBuildAc) {
+                    this.buildAc();
+                }
+            }
+        }
+        List<FoundWord> ans = new ArrayList<>();
+        Node p = root, k = null;
+        for (int i = 0, len = text.length(); i < len; i++) {
+            int ind = text.charAt(i);
+            // 状态转移(沿着fail指针链接的链表，此处区别于DFA模型)
+            while (p != null && p.next.get(ind) == null) {
+                p = p.fail;
+            }
+            if (p == null) {
+                p = root;
+            } else {
+                p = p.next.get(ind);
+            }
+            // 提取结果(沿着fail指针链接的链表，此处区别于DFA模型)
+            k = p;
+            while (k != null) {
+                if (k.flag) {
+                    ans.add(new FoundWord(k.str, k.str, i - k.str.length() + 1, i));
+                    if (!isDensityMatch) {
+                        p = root;
+                        break;
+                    }
+                }
+                k = k.fail;
+            }
+        }
+        return ans;
+    }
 
-		public Node() {
-			this.flag = false;
-			next = new HashMap<>();
-		}
-	}
+
+    private static class Node {
+
+        /**
+         * 当前节点是否是一个单词的结尾
+         */
+        boolean flag;
+        /**
+         * 指向 当前节点匹配失败应该跳转的下个节点
+         */
+        Node fail;
+        /**
+         * 以当前节点结尾的单词
+         */
+        String str;
+        /**
+         * 当前节点的子节点
+         */
+        Map<Integer, Node> next;
+
+        public Node() {
+            this.flag = false;
+            next = new HashMap<>();
+        }
+    }
 }
diff --git a/hutool-core/src/test/java/cn/hutool/core/text/dfa/NFATest.java b/hutool-core/src/test/java/cn/hutool/core/text/dfa/NFATest.java
index e38be87f3..978d1ee23 100644
--- a/hutool-core/src/test/java/cn/hutool/core/text/dfa/NFATest.java
+++ b/hutool-core/src/test/java/cn/hutool/core/text/dfa/NFATest.java
@@ -16,7 +16,7 @@ public class NFATest {
 	public void testFind() {
 		final NFA NFA = new NFA();
 		NFA.insert("say", "her", "he", "she", "shr");
-		NFA.buildAc();
+//		NFA.buildAc();
 
 		final WordTree wordTree = new WordTree();
 		wordTree.addWords("say", "her", "he", "she", "shr");
@@ -53,7 +53,7 @@ public class NFATest {
 	public void testFindNotDensity() {
 		final NFA NFA = new NFA();
 		NFA.insert("say", "her", "he", "she", "shr");
-		NFA.buildAc();
+//		NFA.buildAc();
 
 		final WordTree wordTree = new WordTree();
 		wordTree.addWords("say", "her", "he", "she", "shr");
@@ -89,7 +89,7 @@ public class NFATest {
 		stopWatch.start("automaton_char_buid_find");
 		final NFA NFALocal = new NFA();
 		NFALocal.insert("say", "her", "he", "she", "shr");
-		NFALocal.buildAc();
+//		NFALocal.buildAc();
 		final List<FoundWord> ans1 = NFALocal.find(input);
 		stopWatch.stop();
 
@@ -124,7 +124,7 @@ public class NFATest {
 		stopWatch.start("automaton_cn_build_find");
 		final NFA NFALocal = new NFA();
 		NFALocal.insert("赵", "赵啊", "赵啊三");
-		NFALocal.buildAc();
+//		NFALocal.buildAc();
 
 		final List<FoundWord> result = NFALocal.find(input);
 		stopWatch.stop();
@@ -161,7 +161,7 @@ public class NFATest {
 
 		final NFA NFALocal = new NFA();
 		NFALocal.insert("赵", "赵啊", "赵啊三");
-		NFALocal.buildAc();
+//		NFALocal.buildAc();
 
 		stopWatch.start("automaton_cn_find");
 		final List<FoundWord> result = NFALocal.find(input);
@@ -200,7 +200,7 @@ public class NFATest {
 
 		final NFA NFALocal = new NFA();
 		NFALocal.insert("赵", "赵啊", "赵啊三");
-		NFALocal.buildAc();
+//		NFALocal.buildAc();
 
 		stopWatch.start("automaton_cn_find_not_density");
 		final List<FoundWord> result = NFALocal.find(input, false);