The emergence of social networking services (SNS) provides an opportunity for the application of tag. In this paper, we choose the website based on SNS as the background, and then integrate tag information into collaborative filtering recommendation system. So we proposed collaborative filtering recommendation system based on similarity fusion of tag and rating under the background of SNS. It can help us to reduce the influence of data sparsity on the recommendation accuracy. First, we calculate the user similarity based on Tag information and Rating information respectively, and then get the integrated similarity by the fusion of these two similarities. Finally, Collaborative Filtering can be executed based on this integrated similarity. Experimental results show that the proposed algorithm can improve the accuracy of the recommended.