Scientific monitoring in sports training in colleges and universities is particularly important, which is an important symbol of the scientific level of sports training, and has a role in promoting the ability of athletes and coaches. In this paper, we collect data based on the evaluation index of coach quality and then preprocess the collected data of coach quality and athlete performance into the BP neural network model. The firefly algorithm is used to optimize the prediction performance of the BP neural network, and based on the model, a visualization system for sports training is constructed to display the predicted performance and quality assessment data in real-time. It has been found that the average prediction error of the BP neural network model based on the firefly algorithm is 0.73%, which can be used in sports training. The test scores of athletes in all aspects assisted by the visualization system were significantly better than those of the traditional training group, and the quality assessment scores of coaches in this model were significantly higher than those of the traditional group (P<0.05). This paper forms a systematic method for monitoring college sports training, which can enhance the scientific level of sports training monitoring and help improve the ability of athletes and the quality level of coaches.