Obesity is a major global health crisis, and machine learning techniques show promise for predicting and understanding the factors that contribute to it. This study explores the application of decision tree algorithms with various pruning strategies for predicting obesity. The dataset used in this study includes variables related to demographics, socioeconomic factors, health behaviors, medical information, diet, medication, substance use, and geographic details. Several pruning methodologies are evaluated, including reduced error pruning, minimum sample size pruning, minimum gain sampling pruning, complexity parameter pruning, cross-validation pruning, pessimistic pruning, pre-pruning, and post-pruning with information gain. Results indicate that certain pruning strategies like minimum gain sampling pruning, complexity parameter pruning, and cross-validation pruning significantly improved the prediction accuracy compared to the unpruned model. The findings in this study highlight the importance of proper pruning techniques in developing robust decision tree models for obesity prediction. Overall, this research provides valuable insights into optimizing machine learning approaches for tackling the complex nature of healthcare data.