Different from the artificial classification of the inscription and seal, the computer classification uses image content characteristic as a source of information and is a key work in the digital management. Aiming at the non-standard problems and deficiencies of the existing feature extraction algorithms, a segmentation filtering method based on image entropy is proposed. The texture characteristics of traditional Chinese paintings are extracted by the method combined with the complex network theory, and then a support vector machine is used to classify. Experimental results show that this method can effectively extract textural features and categorize Chinese paintings, and it still has a good performance in the case of non-standard images.