Since 1990s, with an advancement of network technology and the popularization of the Internet, information that people can access has proliferated, thus information recommendation has been investigated as an important issue. Because preference to information recommendation can be different as context that the users are related to, we should consider this context to provide a good service. This paper proposes the recommendation system that considers the preferences of group users in mobile environment and applied the system to recommendation of restaurants. Since mobile environment has plenty of uncertainty, our system have used Bayesian network which showed reliable performance with uncertain input to model individual user's preference. Also, restaurant recommendation mostly considers the preference of group users, so we have used AHP (Analytic Hierarchy Process) of multi-criteria decision making method to get the preference of group users from individual users' preferences. For experiments, we have assumed 10 different situations and compared the proposed method with random recommendation and simple rule-based recommendation. Finally, we have confirmed that the proposed system provides high usability with SUS (System Usability Scale).