Traditional case teaching in international trade is constrained by slow updating, fragmented business materials, uniform task difficulty, and weak connections with changing trade scenarios. Generative artificial intelligence (GenAI) offers a way to transform static cases into adjustable teaching resources. This study adopts literature analysis, instructional design research, and scenario-based case analysis. Taking the export of automated packaging equipment from a Chinese manufacturer to a Malaysian buyer as an illustrative case, it proposes a framework consisting of goal anchoring, constrained generation, teacher verification, layered task delivery, learning-evidence collection, and iterative optimization. The framework generates interconnected materials such as inquiries, quotations, contracts, documentary credits, commercial documents, role instructions, and risk events, while retaining teachers’ control over professional accuracy and pedagogical suitability. A human-AI collaborative quality mechanism is established to address factual accuracy, rule compliance, process consistency, difficulty alignment, and traceability. Rather than claiming experimentally verified learning gains, the paper clarifies the instructional value, operating conditions, and governance boundaries of the model. GenAI should be positioned as a resource co-creation and variation engine rather than an autonomous teacher. The framework provides a feasible reference for improving the timeliness, authenticity, differentiation, and sustainability of case teaching resources in applied international trade courses.