Purpose/Significance The paper systematically reviews and analyzes the development, cutting-edge applications, challenges and alignment issues of large language models (LLMs) in healthcare, so as to provide references for promoting the safe, effective and equitable application. Method/Process Through literature analysis, the development and application of LLMs in healthcare are expounded, the challenges faced are analyzed, and countermeasures are proposed from the technical and social aspects. Result/Conclusion LLMs in healthcare should prioritize truthfulness, safety, interpretability, fairness, privacy and accountability to realize model alignment, meet ethical requirements, and improve patients’ welfare.