The invention discloses a user network behavior prediction method based on vectorization characteristics, and belongs to the technical field of data mining and behavior prediction. According to the invention, the problem of low accuracy of the user network behavior prediction in the existing method is solved. According to the method, behavior transaction division is carried out on the network behaviors of the users by using the user access logs related to the users and according to the relevance characteristics of the network behaviors of the users; the feature extraction is carried out on theuser network behaviors according to the divided network behavior affairs; and finally, the correlation analysis is carried out on the user network behaviors according to the extracted feature vectorsof the user network behaviors, the user network behaviors with strong correlation are merged into one class, and the network behaviors of the users are predicted according to a classification result.According to the method, the user network behaviors with high relevance are combined, so that the accuracy of predicting the user network behaviors is improved. The method can be applied to the prediction of the user network behaviors.