Abstract: This study addresses the challenges of insufficient customer engagement, static weight allocation, and the semantic gap between demand and resource descriptions in complex heavy equipment manufacturing resource matching by proposing a dynamic mapping and recommendation method driven by customer preferences. The approach entails constructing a customer-oriented demand indicator system to capture differentiated preferences, alongside a capability-centric resource evaluation system to characterize enterprise attributes. An integrated mapping model is then developed by fusing Quality Function Deployment(QFD), the Analytic Network Process(ANP), an improved CRITIC method, and a game-theoretic aggregation model, which facilitates the precise translation of customer demands into evaluation criteria. A dual-feedback dynamic update strategy(pre-and post-service) is designed, forming a closed-loop mode-mapping-update framework to ensure sustained system accuracy. A case study demonstrates the method's efficacy in differentiating resource capabilities, enhancing personalized response precision and manufacturer adaptability, thereby offering a practical decision-support tool for networked collaborative manufacturing with significant application potential.