Abstract: ObjectivesIn order to solve the problem of low efficiency in ship equipment fault diagnosis and difficulty in effective information communication with commanders, a solution is proposed that uses natural language interaction to quickly locate the cause of faults and maintenance plans.MethodsBased on the domain oriented design concept, an intelligent question answering system was constructed using a large language model and Retrieval Augmented Generation technology; Then propose a set of document preprocessing methods and comprehensive retrieval strategies to optimize system performance, and design a comprehensive evaluation plan to comprehensively evaluate the system.ResultsExperimental results have shown that compared to basic question answering systems, the optimized question answering system has doubled its ROUGE score, increased its BERTScore score by nearly 30%, increased expert ratings by 1.5 times, and reduced system response time by 95% compared to traditional manual retrieval methods.ConclusionsSimply using natural language to describe the fault phenomenon can quickly locate the cause of the fault and provide maintenance solutions, significantly improving the efficiency of fault diagnosis and providing strong technical support for the rapid recovery of equipment performance of coast guard vessels in complex mission environments, thereby effectively enhancing the combat effectiveness and mission execution capability of the vessels.