Along with the network popularization and shopping in-depth development in China, rational utilization of network information has received attention from a wide variety of areas. Via the understanding and predicting online consumers' needs and preferences in e-commerce, to achieve customer satisfaction and cross-selling effectively is the premise and guarantee to locate targets for online retailers. Based on online consumers' basic information, online shopping information and online behavior information under network environment in e-commerce, this paper proposes a cross-selling index system which fits various characteristics of online retailers.