We propose a new hybrid recommendation algorithm to optimization the cold-start problem with Collaborative Filtering (CF). And we use neighborhood-based collaborative filtering algorithm has obtained great favor due to simplicity, justifiability, and stability. However, when faced with large-scale, sparse, or noise affected data, nearest-neighbor collaborative filtering performs not so well, as the calculation of similarity between user or item pairs is costly and the accuracy of similarity can be easily affected by noise and sparsity. We introduce a new model comprising both In the training stage, user-item and film-item relationships in recommender systems, and describe how to use algorithm generates recommendations for cold-start items based on the preference model. Our experiments model provides a relatively efficient and accurate recommendation technique.