本文设计了一个基于内容的ＥＭＡＩＬ安全拦截系统模型．该模型从用户的角度出发，为了满足不同用户对不同主题的需求，采用层次智能信息滤波算法和有专家指导的相关反馈学习算法，过滤掉不可信度大于阈值的非法 Ｅｍａｉｌ．使收信者甚至意识不到非法Ｅｍａｉｌ的存在，有效地起到了安全卫士的保护作用． This paper designs a content-based EMAIL security interception system model. In order to satisfy different users’ demands on different topics, this model uses the hierarchical intelligent information filtering algorithm and the relevant feedback learning algorithm guided by experts to filter out illegal emails with untrustworthiness greater than the threshold. So that recipients do not even recognize the existence of illegal Email, effectively played a protective role of security guards.