As an independent sport, aerobics is a popular and highly popular sport that integrates group gymnastics, dance, music, fitness, and entertainment. With the continuous development of global fitness fever and happy sports in recent years, the classification of aerobics has become more and more extensive. The development of fitness aerobics and competitive aerobics conforms to the trend of the times. In the course of aerobics teaching, action structure analysis is essential, and the performance of decomposing action image recognition has a direct impact on the effect of aerobics teaching. Therefore, it is important to study an effective aerobics complete set of action structure analysis method. In order to solve the problem that the traditional method is easily affected by the movement rate, lighting conditions, occlusion, complex background, etc., resulting in poor robustness of the recognition results, this paper takes the complete set of aerobics movements as the initial population and then designs the adaptation of the aerobics movements. The degree function is used to realize the structural analysis of aerobics complete sets of movements through the genetic algorithm. First, the video sequence is divided into several segments by using the time energy pyramid, and the human body target for aerobics exercise is extracted by the background subtraction method, and the decomposition actions of aerobics are obtained. By calculating the fitness of different actions, the fitness of the complete set of actions is obtained and then realize the analysis of aerobics complete sets of movements. The structural analysis of aerobics complete sets of movements can not only improve the choreography efficiency of complete sets of movements, but also promote the improvement of the training level of aerobics athletes, and help coaches make relative adjustments and changes in teaching on the basis of the original training of athletes.