Recommender systems have received considerable attention in recent years. Yet with the development of information technology and social media, the risk in revealing private data to service providers has been a growing concern to more and more users. Trade-offs between quality and pri- vacy in recommender systems naturally arise. In this pa- per, we present a privacy preserving recommendation frame- work based on groups. The main idea is to use groups as a natural middleware to preserve users’ privacy. A dis- tributed preference exchange algorithm is proposed to en- sure the anonymity of data, wherein the effective size of the anonymity set asymptotically approaches the group size with time. We construct a hybrid collaborative filtering model based on Markov random walks to provide recom- mendations and predictions to group members. Experimen- tal results on the MovieLens dataset show that our proposed methods outperform the baseline methods, L+ and Item- Rank, two state-of-the-art personalized recommendation al- gorithms, for both recommendation precision and hit rate despite the absence of personal preference information.