Edge computing and artificial intelligence, two new technologies with rapid development, have great potential to empower each other, and have been widely used in various fields, with obvious results. Unlike other discipline-based teaching, physical education (PE) has no fixed classrooms, no fixed teaching materials, and no fixed teaching tasks. Therefore, it is difficult to uniformly evaluate the quality of PE teaching in ordinary colleges. In this paper, the edge computing optimization model is used to optimize the PE quality evaluation model in ordinary colleges. The goal of PE in ordinary colleges and universities is to improve the physical quality of students. Thus the teaching evaluation model in this paper focuses on the physical fitness compliance rate of students' evaluation, and re-weights the proportion of final exam scores. For the monitoring of PE teaching quality in ordinary colleges, this paper designed a PE teaching health system, which allows students to evaluate in the entire learning process, so as to achieve real-time monitoring of teaching quality. It uses the edge computing technology to optimize the system. The experimental results of this paper showed that the PES in general colleges designed in this paper can effectively collect students' evaluations, and can effectively improve the scientific nature of PE teaching goals. At the same time, the edge computing optimization algorithm designed in this design can reduce the network transmission cost of the system by 20% and increase the transmission efficiency by 15%.