In order to further improve the prediction effect of college students’ sports performance, this paper tries to introduce genetic algorithm on the basis of sparrow search algorithm, which can effectively solve the problem that the model is easy to fall into local optimum during parameter search. By applying the selection, crossover, and mutation strategies of GA, the optimized algorithm not only maintains genetic diversity, but may also escape from the local optimum, thus improving the convergence speed. Moreover, this paper constructs a college students’ sports performance prediction system based on the improved sparrow algorithm, and analyzes the performance of the system with experiments. Through the experimental results, it can be seen that compared with other optimization algorithms, this algorithm has stronger global search ability and wider search range, and the prediction accuracy of the model is higher. Therefore, its predicted indexes are superior to other methods, and it can be well applied to the prediction of college students’ sports performances. Finally, by finding the optimal parameters of the model, the prediction accuracy, universality and generalization ability of the model can be improved.