C Programming is a foundational core course for computer-related majors in higher education. It has three cognitive barriers: intricate syntax

the opaque "black box" of compilation

and disconnection between teaching cases and practical application. Traditional teaching methods struggle to cultivate students' metacognitive abilities

while conventional inquiry-based teaching is constrained by two bottlenecks: difficulties in visualizing abstract concepts and high costs of constructing individualized learning pathways. To address these problems

this paper proposes an AI-empowered chain-based inquiry teaching model for C Programming and develops a locally deployed intelligent teaching platform. Taking the "inquiry task chain" as the main teaching thread and "diagnosis-probing-archiving" as the core teaching logic

the platform organizes learning processes through eight-step progressive inquiry activities. It provides dynamic learning scaffolds relying on heuristic probing of large language models and compilation behavior diagnosis

and realizes whole-process evaluation based on a three-dimensional model. A comparative experiment was carried out on two parallel classes in our university. The experimental class achieved significantly higher scores in final examinations than the control class and gained positive outcomes in the development of inquiry competency and learning experience. The results preliminarily verify the feasibility and effectiveness of the proposed model in improving academic performance and fostering students' autonomous inquiry habits.