This paper proposes a shape context shape context algorithm combined with DTW dynamic regularization to obtain the data of the skeleton joint points of the dance posture, and calculate the similarity and key frame time of the joint points by DTW dynamic time regularization matching to complete the quantitative analysis of the dancer’s movement and rhythm. This paper is based on the application of national dance classes in colleges and universities. The corresponding research results can provide theoretical data reference for dance teachers and dancers. The results show that after the experiment, the calorie data consumed by the students in groups A and B, has a significant difference (P=0.01). After 12 weeks of teaching experiments, the situation of all abilities and independent learning conscientiousness of group A was significantly improved. A total of five advantages of AI dance teaching could be obtained through interviews. Therefore, the results of this paper provide new thinking and ways to realize basic dance education.