<p>In today’s high-tech social background, the traditional aerobics teaching model faces some challenges, such as inaccurate movements of students, poor teaching feedback and so on. In order to make aerobic exercise teaching more effective, this paper chooses sensor network and face recognition technology as auxiliary means, and proposes a face recognition technology based on light sensor, which aims to improve the teaching effect by accurately recognizing students’ facial movements. This paper uses light sensors to collect the data of students’ facial movements. Then the collected data is processed through the face recognition algorithm to identify the specific actions of the students. The identified movements were then compared with standard movements to assess the accuracy of the trainees’ movements. Finally, according to the evaluation results, the students are given the corresponding feedback and guidance. The experimental results show that the face recognition technology based on light sensor can accurately recognize the students’ facial movements. The accuracy of students’ movements can be evaluated effectively by comparing with standard movements. The effectiveness of teaching feedback has also been significantly improved.</p>