In analyzing medical data to support clinical decisions, segmentation of health information plays a crucial role. This study presents a comparative analysis of K-Means and K-Medoids algorithms in clustering Medical Examination data. This evaluation is conducted using two main internal approaches, namely Silhouette Score and Davies-Bouldin Index in measuring the quality of separation as well as cohesion between clusters. The experiment involved varying the number of clusters to determine the optimal configuration of each algorithm. The results show that K-Means provides representative performance and is more stable against data complexity, compared to the K-Medoids algorithm which is only optimal in a small number of clusters. Statistical analysis using one-way ANOVA was applied to test the significance of performance differences between algorithms based on the average Silhouette Score value, yielding an F-value of 4.8594 with a P-value of 0.0447. This indicates that the performance difference between the two algorithms is statistically significant at 5% significance rate. This research confirms the K-Means algorithm for segmenting health data with diverse distributions and is expected to serve as a foundation for the development of more efficient health data classification systems in the future.