To improve the quality of sports training and competitions, it is necessary to analyze sports videos. Therefore, this paper builds a sports video analysis system based on image recognition technology. To improve the data fusion effect of the system, this article starts from the time and space scale and frequency scale of the dynamic image to comprehensively use the technologies of motion estimation, multi-scale transformation, and visual attention to realize the comprehensive analysis of the dynamic image. Moreover, this paper constructs a dynamic image analysis method and applies it to dynamic image fusion to improve the quality of the fused dynamic image. In addition, this article combines the needs of sports image analysis to construct system function modules and compares the input video action decomposition with the standard database to correct the sports action. Finally, this article designs experiments to verify the performance of the system. The research results show that the system constructed in this paper has a certain effect.