To address the challenge of heavy reliance on manual effort in remanufacturing process decision-making due to knowledge with heterogeneous forms and low reuse efficiency, an intelligent decision-making approach driven by explicit and implicit knowledge-augmented large language model (LLM) is proposed. First, explicit and implicit knowledge extraction methods are introduced based on vector embedding and parametric modeling of decision objectives, respectively, constructing a remanufacturing process knowledge base comprising a process vector library and a process function library. Second, a retrieval-augmented generation (RAG) strategy is proposed to drive the LLM in extracting remanufacturing process knowledge from the process vector library, generating a set of feasible remanufacturing process schemes. Then, a function-learning mechanism is established, allowing the LLM to invoke formulas from the process function library to calculate decision objective values and determine the optimal remanufacturing process scheme. Finally, the feasibility of the proposed approach is validated through a case study on the remanufacturing of an automotive engine worm gear. The results show that the proposed approach not only significantly reduces reliance on the experience of process engineers but also accurately generates remanufacturing process schemes and facilitates decision-making, which greatly enhances the automation level and reliability of remanufacturing process decision-making.