Abstract: In order to obtain high efficient hypertext classification algorithms, a new coordinate hypertext classification algorithm is proposed. The k NN, naive Bayes and similarity are also introduced to classify the hypertext, and experiments are conducted to compare the three classifiers’ effects. Finally, a high effective hypertext classifier is obtained. Now, this classification method has been used in two new systems: The intelligent search engine system （WebSearch） and the intelligent software assistant system （WebSoft）.