Due to the rapid growth of the World Wide Web in recent years, extensive research on modeling the behavior of users on the Web site was conducted. In this direction, Web usage mining with the aim of obtaining navigational behavior of web users, has been used by many researchers. Generally, Web users may exhibit different behaviours proportional to their information needs and intended tasks when they are visiting a Web site. All the navigational behavior of Web users in the web access log files is traceable. One of the techniques utilized in web usage mining is the clustering of web users. In clustering techniques, users who have similar navigational behavior are in a cluster. Each cluster is led to the creation of user profiles that are used in web applications such as web prefetching and the caching. The conventional web usage mining techniques for clustering web user can discover usage patterns directly, do not generally provide the ability to automatically characterize or quantify the unobservable factors that lead to common navigational patterns. Therefore, it is necessary to develop techniques that can automatically identify the users’ underlying navigational objectives and to discover hidden semantic relationships among users as well as between users and Web objects. In this project, we propose an approach based on probabilistic latent semantic analysis to discover such intrinsic characteristics of Web users’ activities. The proposed method, obtained latent factors for clustering user navigation patterns used to create user profiles. The clustering results will be used to predict and prefetch web requests for grouped users. The usability and superiority of the proposed web user clustering approach through experiments on a real Web log file will be displayed. The clustering and prefetching tasks are evaluated by comparison with previous studies demonstrating better clustering performance and higher prefetching accuracy. Keywords Web usage mining, Web user clustering, Probabilistic latent semantic analysis, Web prefetching