Purpose/Significance To review the application progress of the integration of knowledge graph (KG) and large language model (LLM) in the intelligent health management of diabetes, and to promote the transformation of diabetes health management from an experience-driven mode to a data and knowledge-driven mode. Method/Process The application research of the integration of KG and LLM in diabetes health management is systematically sorted out, and the collaborative mechanism is analyzed. Combining typical cases at home and abroad, its application in intelligent question answering, personalized intervention and complication prediction of diabetes is expounded. Result/Conclusion The deep integration of KG and LLM significantly enhances the accuracy and efficiency of diabetes health management. In the future, continuous improvement should be made in four aspects:privacy protection, data coverage and interpretability, cross-disciplinary knowledge integration.