<p>With the continuous development of Internet technology, the online teaching system of colleges and universities, as an innovative teaching method, has been widely concerned. However, the online teaching system often lacks real-time and interactive, which limits the communication and interaction between students and teachers. The introduction of sensor technology can effectively compensate for this shortcoming, providing a more personalized and high-quality online teaching experience. The aim of this research is to design a sensor-based interactive speech recognition system to increase the real-time and interactive nature of online teaching system. Through the sensor collection of students’ movement and environment information, combined with voice recognition technology, the sensor collection of students’ movement data and voice data are integrated with the teaching system to achieve real-time teaching feedback and personalized guidance, so as to achieve real-time interaction and feedback between students and teachers, and provide a more flexible and intelligent online teaching environment. Through the design and implementation of interactive speech recognition system based on sensor, remarkable results have been achieved in the online teaching system of colleges and universities. Students can ask, answer and interact through voice, and the system can recognize and transcribe the voice content in real time, providing real-time feedback and guidance, improving the teaching effect and learning experience.</p>