Digital Signal Processing and Artificial Intelligence for Automatic Learning Ning Fu, Xia Peng 3 3 2024 https://doi.org/10.6025/dspaial/2024/3/3/97-103 https://www.dline.info/dspai/fulltext/v3n3/dspaiv3n3_3.pdf This paper is based on the dance movement recognition model of 3D Convolutional Neural Networks (CNNs) and aims to explore the feasibility of applying intelligent technology to dance teaching. By investigating the current status of intelligent technology in dance teaching both domestically and internationally, the advantages of 3D CNNs in dance movement recognition are analyzed. In response to the needfor dance movement recognition, a model based on 3D CNNs is designed and validated using the MSRAction3D dataset. The experimental results show that the model achieves high recognition accuracy in 20 dance movement categories, proving its potential application in dance teaching. This model can provide dance teachers with accurate movement recognition and personalized guidance, improving teaching effectiveness.