To solve the cold start problem of new movies, a collaborative filtering recommendation algorithm (IW-CFW) based on BP neural network is proposed. In this algorithm, the similarity of movies is fused by BP neural network, and the BP network model is optimized and adjusted according to the error between the predicted score and the true score, and the final prediction model is obtained. By comparing the Mean Absolute Error (MAE) and Root Mean Square Error (RMSE) of the proposed algorithm with the other three algorithms on the Movielens and Movie-little datasets, the experimental results show that the proposed algorithm can effectively solve the cold start problem of new movies and produce more accurate recommendation results.