With the proliferation of fake calligraphy, how to effectively identify calligraphy works has attracted more and more experts. The current appraisal of fake calligraphy mainly relies on the subjective judgment of experienced experts, with large uncertainty and high appraisal costs. With the development of digital image technology and deep learning models, the use of computer technology to identify calligraphy fakes has become a feasible option. This paper proposes a calligraphic work recognition method based on deep learning. In view of the diversity of calligraphy works and the difficulty of sample collection, this article only selects some of the regular scripts by Yan Zhenqing, a famous calligraphy master, as the identification object at this stage. The research includes six aspects: the collection of genuine and counterfeit data sets, the selection of identification character sets, the preprocessing of calligraphy images, word segmentation, single word neural network training, and calligraphy authenticity identification. Finally, a complete scheme programs is provided to identify calligraphic works. The test results show that the scheme proposed in this paper can effectively extract the features of the Chinese character, and can correctly judge the authenticity of the work.