Purpose/Significance To investigate the application of large language model(LLM)in supporting clinical diagnosis and treatment，knowledge acquisition，and competency development in obstetrics and gynecology. Method/Process Based on the general LLM and combined with desensitized real case data，an obstetrics and gynecology medical knowledge base and a physician assistantagent are constructed by using retrieval-augmented generation(RAG)technology and verified in the clinical application environment. Result/Conclusion This agent significantly reduces the workload of physicians in case preparation and knowledge retrieval，and at the same time lowers the rate of generating“hallucinations”.