A novel P2P traffic identification method based on the combination of naive Bayes and decision tables is proposed, which uses Fast Correlation-Based Filter (FCBF) algorithm to extract P2P flow characteristics, and utilises six DTNB (combination of naive Bayes and decision tables) combined with dynamic weighted integration method to set up a P2P flow detection model. Through experimental comparison between this proposed model and traditional methods, such as single DTNB, decision tree and naive Bayes, we find that the proposed method has a better P2P traffic identification accuracy and stability.