This paper analyzes the relationship between user portrait and e-commerce, and probes into the construction strategy of user portrait system. From the aspects of precision marketing, user statistics, data mining, building intelligent recommendation system, marketing effect evaluation, improving the operation mode of products and services, user satisfaction management, private customization of products and services, etc., with the continuous growth of user data, the scope and dimensions of user portrait model will be more diverse. According to the data of a shopping mall, multi-dimensional labels such as rule matching, statistics and data mining are implemented to form user portraits. Suggestions are put forward for the development of e-commerce user portrait system.