The development of information technology is accelerating the development of network multimedia-assisted teaching modes. In this regard, this paper studies the influence of multimedia-assisted sports education on the teaching of tennis classes in colleges and universities. The first step is to analyze the applicable fields and characteristics of network multimedia CAI. Then, the TLBO algorithm focuses on analyzing the “teaching” and “learning” phases of tennis teaching and finds that the TLBO algorithm has defects. To overcome the local optimal situation of the TLBO algorithm, a grouping mechanism is implemented to optimize the TLBO algorithm and maintain the diversity of the population. Finally, the influence of multimedia-assisted teaching was studied in terms of learning motivation, physical fitness, and learning effect. From the perspective of liking tennis sport, the number of experimental classes is 20, which is 15 more than the number of control classes. The mean score of the experimental class in the forehand hitting spinning ball test was 19 points higher than that of the control class.