With the in-depth development of artificial intelligence in the field of education and teaching, courses on statistical forecasting and decision-making have revealed various issues such as outdated methods and techniques in adapting to the demands and objectives of the AI era. In response, this paper constructs a heuristic inquiry-based teaching model for the course, divided into three stages: pre-class, in-class, and post-class. It explores how AI can empower each teaching stage as well as the teaching methods on both the teacher and student sides, and provides a lesson design for the topic “stochastic time series forecasting.“ To scientifically evaluate the effectiveness and feasibility of the teaching model, two classes with similar learning conditions were selected for a teaching practice: one as a control group following the traditional model, and the other as an experimental group using the new teaching model and AI-assisted instruction. Analysis of test data and survey responses from this lesson and the entire semester shows that the experimental class using AI-empowered heuristic inquiry-based teaching performed better in terms of both academic achievement and critical thinking skills compared to the control class, indicating the feasibility of the teaching model designed in this study.