Abstract: There are there tough problems in building a personalized search engine system：first collecting user information，second auto-refreshing database and retrieval algorithm utilizing collected information.To solve the three problems，this paper creatively promotes user-tracing method based on Ajax，strategy of refreshing user-action database by session and personalized information retrieval model based on VSM plus user-document.Finally it describes the design and implementation of whole testing personalized system.

摘要： 针对个性化搜索的三个关键问题：用户信息搜集，用户信息库的动态更新与个性化检索算法，探索性地提出了基于Ajax用户行为跟踪方案，以会话为单位动态更新用户行为信息库策略与加入用户文档的向量空间检索模型，在此基础上设计并实现了个性化搜索引擎实验系统。