Purpose/Significance The anonymization algorithm based on clustering has the characteristics of high flexibility, wide application range, and the ability to retain more information of the original data. Reasonable use of clustering-based anonymization algorithm for anonymization can obtain high-quality medical data that meets the needs of privacy protection. Method/Process Through literature research and comparative analysis, the study sorts out the key technologies of clustering-based anonymization algorithm for medical data sharing, summarizes the main process of such algorithm and the related privacy models, including representative traditional privacy models and personalized privacy models, and analyzes the advantages and disadvantages of representative clustering-based anonymization algorithm. Result/Conclusion The type of clustering-based anonymization algorithms should be reasonably selected, the model of clustering-based anonymization algorithms should be flexibly improved, and the research and development of clustering-based anonymization tools should be heavily invested. These measures can promote safe, convenient and higher quality sharing of medical data.