In the era of swift advancements in information technology, the domain of physical education (PE) at colleges and universities stands at the cusp of transformative change, necessitating innovative approaches to teaching methodologies. This paper delves into integrating information fusion technology within PE to refine instructional strategies and bolster student physical health. Through the employment of surface electromyography (EMG) signal collection, Analysis of motion imagery features, and a comprehensive approach to multi-information fusion, this investigation analyzes the EMG data, heart rates, and motion images of 60 students. The findings reveal that information fusion technology serves as an effective tool for monitoring exercise states and physiological metrics in real-time. A notable aspect of the research is using a Dual Stream CNN-ReliefF algorithm, enhancing the precision of interpreting human motion intentions. The study confirms the method’s accuracy and reliability, as the monitored average heart rate, EMG signals, blood pressure, and body temperature align with readings from specialized medical devices. Consequently, adopting information fusion technology in PE teaching significantly enhances educational methodologies and outcomes. This research underscores the potential of technology-driven solutions in redefining physical education, fostering a more effective and scientifically informed learning environment.