Abstract: Information active push is the trend of personalized search engine.How to get user interests and classify the users of the same interest are important.Support Vector Machine(SVM) is widely used,while it faces some challenges,including classification of large-scale datasets and information leakage.In order to solve the above problems,User Classification Method based on Large-scale Information(UCM) is proposed in this paper,it can effectively solve the large-scale information classification and preserve user information during classification.