Smart teaching refers to the in-depth use of modern information technology to promote the process of education, which is characterized by using digital, network, intelligent and multimedia technologies. BOPPPS teaching model is a new type of student-centered teaching model. This teaching model is widely used around the world. The traditional BOPPPS teaching model makes it difficult to implement personalized teaching in the classroom. The new development of artificial intelligence technology provides new method for smart teaching, especially the AIGC Large Language Model represented by ChatGPT. This paper introduces AIGC technology into personalized teaching, and study the application of AIGC in various aspects of BOPPPS teaching model, by taking the design of smart teaching course of video surveillance as an example. In particular, AIGC is utilized for Pre-assessment. It is proposed to use an agent as a mediator between students and the AIGC large language model to test the students individually, and to design personalized test contents for personalized feedback. Comparatively better results were obtained in the actual teaching process, which can provide a reference for other smart teaching researchers.