People's life and habits have changed with the increasing usage of mobile devices and applications. In response to this market trends, many companies have joined the development of apps. Several apps have been developed that allow users to search product information. This kind of apps can improve user's experience if they can provide recommendations. In this study, we use the behavior information of product comparing records from a car app to create car recommendation system. The association rule analysis was adopted as the miming method. Through the empirical analysis, we develop recommendation rules for car information app. The conclusions and managerial implications will also be discussed.