The main ventilation fan ensures the safety of the working environment and the normal operation of the mine, making it the core of the modern coal mine ventilation system. Therefore, enhancing the intelligence level of main ventilation fan management is an inevitable trend. Currently, research on intelligent monitoring of main ventilation fans is relatively mature, while studies on intelligent Q&A systems are scarce. To address the “hallucination” problem in domain-specific Q&A systems built using large language models (LLMs), this paper proposes a method based on knowledge graph-enhanced LLM. Building on the existing knowledge graph in the field of main ventilation fans, data preprocessing is conducted using NLP technology to generate datasets, and the GLM-4-9B model is fine-tuned with LoRA. Combined with a front-end module, an intelligent Q&A system for ventilation fans is constructed. Experimental results show that this system exhibits excellent accuracy in Q&A within the main ventilation fan domain, especially in handling complex questions, effectively mitigating the “hallucination” problem of LLMs and improving the management level of main ventilation fans and the efficiency of personnel.