This study aims to analyze inpatient data using the K-Means Clustering method on a simulated dataset. The dataset includes various patient-related attributes such as age, billing amount, length of stay, medical condition, and type of admission. Several preprocessing steps were applied, including date conversion, duration calculation, numerical normalization, and one-hot encoding for categorical attributes. The Elbow Method was used to determine the optimal number of clusters, and clustering quality was evaluated using both the Silhouette Score and Davies-Bouldin Index. The analysis results show that the patients can be segmented into three major clusters, each exhibiting distinct characteristics—for example, younger patients with short and low-cost stays, and elderly patients with prolonged and more expensive hospitalizations. The resulting Silhouette Score of 0.14 and Davies-Bouldin Index of 1.74 reflect a moderate clustering performance, yet the model remains informative and meaningful. These clusters provide actionable insights that hospitals can use to optimize their service strategies, improve resource allocation, and enhance operational efficiency. Moreover, the study illustrates the practical application of unsupervised learning techniques in healthcare settings, contributing to data-driven decision-making practices and offering a foundation for further research into patient segmentation.