Fuzzy clustering analysis is a clustering algorithm based on function best practices, technology and optimal cost function using calculus. Fuzzy clustering, each sample is no longer belong to a class, but belong to a certain degree of membership of each class. In this paper, Web log sequential pattern mining knowledge gained, and visitors have the same browsing mode access to cutting the interaction of users with the Web information space. The paper presents analysis of Web log data mining based on improved fuzzy clustering algorithm. The experiment demonstrates the improved algorithm has better scalability.