As a frontier concept of current basic education, subject literacy indicates that our country has established the direction of curriculum reform aiming at developing students' core literacy. The core subject literacy is not only the basic subject literacy gradually developed by students in the course of subject learning, but also the comprehensive factors such as character and ability gradually formed by students in extracurricular social practice activities, which is an important indicator of the new curriculum education reform. In view of the fact that the special teaching connotation of high school physical education still needs to be clarified, the evaluation methods of teaching effect still need to be enriched, and the teaching reform theory needs to be further improved. Focus on cultivating college students' lifelong sports awareness, mastering sports knowledge and skills, and developing their physical fitness, good sports behavior and sports moral quality healthily. According to the characteristics of physical education learning evaluation, the fuzzy neural network is selected to model, and the physical education teaching evaluation system is constructed. The adaptive fuzzy neural system is used to process the data, thus realizing the intelligence and automation of physical education teaching evaluation. Finally, the artificial neural network and adaptive fuzzy neural system are used in the evaluation and modeling of physical education teaching, and the experimental simulation comparison and performance analysis of the two models show that the model has good evaluation ability.