A hybrid graph model for personalized recommendation, which is based on small world network and Bayesian network, is presented. Small world network has a good property in clustering and Bayesian network is compatible for probability inference. The hybrid graph model consists of two layers. One is user’s layer for representing users or customers and the other is merchandise’s layer for representing goods or products. Small world network describes the relationships among the nodes of users in lower layer. The implications among nodes of merchandises are represented by Bayesian network in higher layer. Directed arcs denote the tendency of nodes between user’s layer and merchandise’s layer. This paper also introduces several algorithms for clustering based on small world network, structure learning and parameter learning of Bayesian network, and recommended algorithm based this model. The experimentation shows that the model be accomplished to represent the relationships from user to user, merchandise to merchandise, and user to merchandise. The experimental results show that the hybrid graph model has a good performance in personalized recommendation.