With the rapid development of the interdisciplinary area of artificial intelligence and medicine, large language model (LLM) has been widely used in the fields such as diagnosis and treatment, medicine, and healthcare. LLM has unique advantages in the field of traditional Chinese medicine (TCM), such as high consistency with the "Four Diagnostic Methods", perfect combination of natural language and self-supervised learning in TCM, the ability to adapt to the characteristics TCM formulas, and the assistance in TCM diagnosis and treatment. At present, various LLM models have been developed, including the "Qihuang Ask Big Model" and the Digital Traditional Chinese Medicine Big Model "GLM-130B", but they still face challenges such as value mismatch and medical abuse, increased demand for interpretability, lack of advanced technology, and domestic policy access. This article reviews the evolution of LLM, its unique advantages and applications in the field of TCM, the problems and challenges, and the future development trends, in order to providereference for the further promotion of LLM in traditional medicine.