The rapid iteration and deep application of artificial intelligence (AI) technologies are driving a comprehensive transformation in higher education—shifting from traditional lecture-based instruction toward digital, intelligent, personalized, and inquiry-based teaching paradigms. This paper takes the core civil engineering course “Structural Mechanics-1” as a practical vehicle to systematically explore the underlying logic, implementation pathways, and practical outcomes of AI-driven pedagogical paradigm transformation. Focusing on three core dimensions—innovation in teaching methods, reconstruction of teaching models, and optimization of teaching systems—the study constructs an integrated AI teaching framework comprising “knowledge graph–intelligent question bank–virtual inquiry scenarios–process-based evaluation.” By developing a knowledge graph for the “Structural Mechanics-1” course, creating a smart teaching platform, building virtual inquiry learning environments, and implementing a full-process intelligent evaluation covering pre-class, in-class, and post-class stages, this research addresses persistent issues in traditional teaching, such as “teacher-centered lecturing, passive learning, coarse assessment, and weak interaction.” The transformation facilitates a shift in the teacher’s role from knowledge transmitter to learning guide, a shift in students from passive recipients to active inquirers, and a shift in teaching resources from static presentation to dynamic adaptation. Practical results demonstrate that AI-empowered pedagogical paradigm transformation can significantly enhance students’ learning interest, self-directed learning abilities, and engineering innovation capacities. This study provides a replicable and scalable practical paradigm and theoretical reference for the development of first-class courses in civil engineering, the cultivation of emerging engineering talents, and the digital transformation of higher education.