Abstract: The traditional recommendation strategy that aimed at all users adopting the same recommendation method could not satisfy current E-commerce enterprises and users'needs.The advanced one was that centered in users and implemented information recommendation on the base of the fully mined users'value.This paper divided E-commerce users into new user,advanced user and ordinary user.It proposed the E-commerce intelligent recommendation strategy based on user segmentation.According to the difference of users,the paper designed different recommendation methods to guarantee enterprises and users'profit.The experimental results indicate that the proposed recommendation strategy and the designed recommendation methods are efficient.