Objective To analyze the correlation between the performance of military academy students’ 5000-meter running and their exercise physiology indicators associated with 5000-meter running so as to make further efforts to arrange training projects scientifically and provide a theoretical basis for military academy to improve the training quality of 5000-meter running. Methods 5000-meter running performance and 12 exercise physiological indexes grouped into four categories (including anthropometry, cardiorespiratory function, velocity fitness and strength fitness) of 58 students from a military academy was measured by cluster sampling method. Gray correlation model was established to analyze the collected data. Then BP neural network model was used to verify the evaluation results. Results According to the order, the five main factors influencing the performance of 5000-meter running were VO_2max, heart rate of 5000-meter running at fixed workload, performance of 2000-meter running, fat rate and FVC index.BP neural network model was able to accurately evaluate the endurance performance of 5000-meter running with VO_2max and heart rate of 5000-meter running at fixed workload. Conclusion The factors screened by gray correlation model are reasonable and credible. Suggest military academy should consider arranging some training projects which contribute to improvement of cardiopulmonary function and drop of fat rate. By this way, the blindness of training is reduced so as to improve the training quality of 5000-meter running.