As a core component of national strategies, interdisciplinary integration is critical to cultivating innovative interdisciplinary talents and promoting scientific and technological progress. Medical-engineering talent training has become a key focus of higher education reform and has been widely explored by domestic universities. However, its systematic construction still faces prominent problems, including insufficient top-level curriculum design, underdeveloped industry-education integration platforms, inadequate faculty collaboration mechanisms, and incomplete diversified talent evaluation systems. By reviewing typical domestic and foreign medical-engineering talent training models and integrating cutting-edge artificial intelligence technologies, this study explores innovative mechanisms and practical training paths for the intelligent era. It proposes four optimization approaches: intelligent teaching content generation driven by multimodal large models, practical teaching innovation supported by AI virtual simulation and intelligent design, personalized teaching realized through large model-based academic diagnosis and adaptive learning, and an integrated talent evaluation system constructed via multi-dimensional data and intelligent algorithms. This research provides theoretical references and practical guidance for building a high-level medical-engineering interdisciplinary talent training system in Chinese universities.