The rapid development of tourism and Internet leads to the increasingly serious problem of tourism information overload. Therefore, tourism recommendation method is very important to solve the problem of information overload. Traditional recommendation algorithms only calculate similarity between users and items based on the score and basic attributes, behavioral needs and comments with tourist emotional factors are ignored. In this paper, Convolutional Neural Network(CNN) is utilized to classify the feature extraction of text comments, Pearson similarity formula is adopted to calculate similar user groups, and Mean Absolute Error(MAE) is employed to evaluate the error of the results. Compared with the traditional collaborative filtering method, the experimental results show that the proposed model can effectively reduce the prediction error.