In this paper, we develop an information recommendation system based on information edited by various users. This system targets a social bookmark service. Our method for finding appropriate information is to gather each user?s bookmark data on the Web and develop a system that analyzes the user?s association. Assuming that users that have similar interests also have information associated to those interests, we describe user co-citation relationships as a network. We extract the sub-graph as the community in which each user centers from this network. And we develop a web service designed to automatically recommend new information regarding these interests. That service has a list of users who share similar preferences with the specified user and a system to present recommended information to the user.