The widespread application of BIM technology in rail transit construction has significantly enhanced the efficiency of rail transit project management. Nevertheless, as project scale and complexity escalate, traditional BIM technology management methods encounter challenges in information integration and decision support. To address this issue, this paper proposes a large-model intelligent BIM technology management knowledge question-and-answer system based on fine-tuning of Llama3.1. This system leverages data processing, model fine-tuning, retrieval augmentation, and other technical means to establish an intelligent system capable of swiftly and accurately responding to BIM management demands. Experimental results demonstrate that the fine-tuned and database-augmented Llama3-8B model outperforms other models in BIM skill proficiency exams, achieving an accuracy rate of 77.82% and receiving high evaluations in terms of fluency, compliance, professionalism, and accuracy. This system notably elevates the level of intelligence in BIM management, presenting vast potential for widespread application.