This study addresses key contradictions within the vocational education curriculum system - namely technological generation gaps, fragmentation of the curriculum resource ecosystem, and insufficient dynamic adaptability. We establish an optimization framework for vocational education large models based on a "diverse data inputs-computing power support-algorithm-driven"approach. By constructing a four-dimensional curriculum optimization model (encompassing objectives, content, strategies, and evaluation) at the theoretical level, this framework promotes collaborative resource co-construction among multiple stakeholders, the development of multimodal intelligent teaching systems, the establishment of dynamic evaluation mechanisms, and the enhancement of teachers' digital literacy in practice, thereby forming a systematic solution. Empirical studies in public security vocational education demonstrate that this model improves curriculum update timeliness by 96.1%, increases the utilization rate of practical training resources by 150%, enhances real-case scenario matching by 69.1%, and boosts internship unit recognition by 30.6%. These results verify the feasibility of curriculum optimization driven by vocational education large models and provide theoretical support and practical paradigms for the digital transformation of vocational education. This research holds significant value for deepening industry-education integration and cultivating comprehensive technical and skilled talents. © 2025 IEEE.
Author Keywords
Index Keywords
