Abstract: Under the background of emerging engineering education, aiming to cultivate application-oriented undergraduate talents in the new energy vehicle direction of vehicle engineering, this paper focuses on microcontroller embedded system teaching, takes microcontroller control system theory as the theoretical foundation, and introduces artificial intelligence convolutional neural network theory. With the aid of AI hardware modules such as Canaan image recognition and iFLYTEK speech recognition, a project-based experimental teaching platform is constructed, using a smart car as the carrier. Through the establishment of a phased and multi-level experimental teaching model and diversified assessment system, the platform achieves the teaching objectives of microcontroller embedded system course, improves the teaching efficiency of artificial intelligence technology, and enhances students’ understanding and application skills of AI-empowered microcontroller embedded system course knowledge.