Information of network grows up fast, and there is an important thing providing user a tool which could search information quickly. In order to achieve this purpose and we must track and analyze user behavior of network. We apply data mining approach which is used to accurately capture user behavior of network traffic. This research is used Petri-net method to extract the user behavior accurately, and it offers path of user behavior. And we use converted weight matrix method to construct prediction page set determining and prediction model, and it has flexibility to make management and access of the database more convenient. The experimental results showed that improve site of website and predict path of user behavior efficiently.