Generally, physical education (PE) is one of the fundamental ability developments of human healthiness. In society, there are many challenges to enlarge the performance in Chinese physical activities. Also, the involvement of 5G communication network technology becomes an advent in physical activity expansion day-by-day in China. Physical activity can help to improve the Chinese's mental ability, self-concept, aim orientation and also prevent states like depression, anxiety, etc. Physical activity without education is like having a body without a soul. There is no debate about the importance of physical education and various forms of exercise in the overall educational system. The PE teaching performance of Chinese developed by several network technologies such as artificial neural networks, IoT, machine learning, This paper approaches the development and refinement in physical education teaching based on 5G network technology to obtain everlasting data without termination. Firstly the sports dataset can be pre-processed using a stacked denoising autoencoder (SDAE). Gaussian Mixture Model (GMM) is utilized for the feature extraction process. A random forest approach (RFA) can be used in the selection of the features. The classification is done by using a CNN-based upgraded classifier. An efficient data allocation (EDA) algorithm is used for storing data in a 5G network. The stored data can be enhanced by using the glowworm swarm optimization (GSO) technique. The result can be simulated in the MATLAB software tool. Finally, the performances such as accuracy, sensitivity, specificity, and memory utilization of various classifiers are analyzed in this paper.