本文提出了一种快速汉语自动分词算法。其主要思想是利用汉语中两字词占７５％的统计规律，提出了两字词根和两字词簇的概念。算法把三音节以上的词用两字词簇来压缩处理，也就是把长词的扫描范围限定在词汇量很小的词簇内，从而不仅提高了分词速度，而且彻底解决了传统最大匹配分词算法中最大匹配词长的设定问题。另外，本文还提出了用两字词簇快速检测交叉歧义的算法。本文的分词算法简洁、速度快、易于实现 This paper presents a fast Chinese automatic segmentation algorithm. Its main idea is to make use of the statistical law of 75% of Chinese words, and put forward the concept of two-word roots and two-word clusters. The algorithm compresses the words above three syllables by two word clusters, that is, limits the scanning of the long words to the clusters with small vocabulary, which not only improves the speed of word segmentation, but also completely resolves the traditional maximum matching word segmentation The problem of setting the maximum matching word length in the algorithm. In addition, this paper also proposed an algorithm for rapidly detecting cross-ambiguity with two-word cluster. The segmentation algorithm in this paper is concise, fast and easy to implement