The application of artificial intelligence in physical education (AIPE) has provided new ways to improve learning and teaching activities in physical classes. However, literature reviews that provide a systematic review and analysis of AIPE are limited. To address this gap, this study provided an overview of AIPE-related empirical research. Specifically, it examined the general state of AIPE, algorithms used for AIPE, and the impact and challenges of AIPE. Following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses, 130 empirical studies related to AIPE were included in the final synthesis. The findings of this study demonstrated that numerous studies have explored the use of AI technologies to enhance physical education classes and training processes. These technologies have been widely employed in athletic performance analysis, health monitoring, and personalized training. AIPE offered great potential for providing personalized instruction, real-time feedback and assessment, and diverse learning environments. However, the use of AI technology poses challenges, including technical reliability and accuracy, privacy and security issues, as well as technical training and teacher support. These findings provide insights for future research on AIPE.