Abstract: With the rapid development of e-commerce，the importance of mining and predicting user’s navigation patterns grow larger than before.As an important task of Web usage mining，mining users’ navigation patterns is the fundamental approach for generating recommendations.But the users’ interests are changeable，and it is difficult to track the exact user navigation patterns.An ant colony approach for this problem is proposed in this paper.In this approach，the web users is been considered as artificial ants，and the ant colony behavior is used as a metaphor to guide user’s choice in the Web site.Firstly，a user navigation model is built，based on ant colony behavior and Web logs.Secondly，an algorithm for mining users preferred navigation patterns dynamically is designed.The experimental results proved the effectiveness of our approach.