The invention discloses a calligraphy style recognition method based on a convolutional neural network. The method comprises the following steps: obtaining and preprocessing calligraphy works to obtain calligraphy samples; dividing the calligraphy samples into a training sample set and a verification sample set; inputting the training sample set to the model classifier and carrying out iterative training, and verifying the model classifier by the verification sample set; inputting the calligraphy samples into the trained model classifier to complete the recognition of the calligraphy style. The recognition method of the invention simplifies the process flow of the calligraphy style recognition, integrates the feature extraction of the calligraphy image and the style recognition, and has the advantages of fast recognition speed and high recognition accuracy. The model classifier has fewer parameters, less training time, faster convergence speed and higher recognition accuracy. The method can continuously increase the type of style, and exhibits good robustness and expandability.