Tuberculosis (TB) remains a global health threat, especially in low- and middle-income countries like Thailand, where rural areas have limited healthcare resources. This study uses a machine learning–based clustering framework to stratify TB patients and screening data in Ngao District, Lampang Province. It combines data sources—including a registry of 899 TB cases and 15,318 chest X-rays—and addresses challenges like data heterogeneity, missing values, and inconsistent formats through cleaning, imputation, and feature engineering. Using the k-means algorithm, evaluation metrics such as Silhouette Score, Davies–Bouldin Index, and Calinski–Harabasz Index confirmed three optimal clusters. Each cluster shows distinct demographic, clinical, and epidemiological features, highlighting TB risk diversity. The findings demonstrate machine learning’s potential to support targeted public health interventions, optimize resources, and improve TB prevention and control. This scalable framework offers insights into other regions facing similar infectious disease challenges, integrating AI into public health systems.