Innovative Practice and Exploration of Large-Model-Empowered Project-Based Teaching in Higher Vocational Education

Abstract

With the rapid development of generative artificial intelligence represented by large language models, vocational education is encountering critical opportunities for teaching model transformation. At present, project‑based teaching for software technology majors in higher vocational colleges faces prominent challenges, including insufficient differentiated instruction, lack of personalized guidance, and disconnection between teaching content and industrial demands. Taking the course Fundamentals of HTML5 Application Development as a practical carrier and drawing on the reform practice of “Large Model + Smart Shipping” in our college, this paper systematically elaborates the implementation path of large‑model‑empowered project‑based teaching. This study proposes an innovative teaching mode integrating a three‑tier hierarchical project system, human‑AI collaborative learning guidance, and multi‑dimensional process‑oriented assessment. By constructing a course‑specific prompt library, a hierarchical project case library and an intelligent auxiliary evaluation system, full‑chain teaching empowerment of large models is realized throughout pre‑class, in‑class, after‑class and assessment phases. Practical results show that full‑process empowerment by large models can effectively address pain points such as inadequate hierarchical adaptation and shortage of personalized guidance in project‑based teaching. It provides a replicable practical paradigm for the digital and intelligent transformation of curriculum teaching for software technology majors in higher vocational colleges.