Purpose/Significance To construct the clustering model of hospital clinical department performance using different algorithms , and to discuss the optimal clustering algorithm and the characteristics of clinical department performance distribution. Method/Process 3 clustering algorithms, including k-means, k-medoids and GMM, are used to analyze the performance data of clinical departments from June 2019 to June 2024, and the silhouette score is used to evaluate the model results, and the performance distribution characteristics are visualized by principal component analysis (PCA) technology. Result/Conclusion The k-means clustering algorithm has the best effect in the model of hospital performance data analysis, and the clinical department performance presents 3 major categories of distribution, which provides references for the operation performance management of hospital departments.