This study focuses on early detection and prevention strategies in smart healthcare systems using diverse health datasets. A gap in healthcare systems is the challenge of timely and accurate disease detection, particularly for chronic conditions like diabetes, cardiovascular diseases, and cancer. The goal is to leverage dataset, the Kaggle Diabetes to develop predictive models for early diagnosis and intervention. The methodology includes data pre-processing, model training, and evaluation using Logistic Regression and Decision Trees. Key steps involved cleaning the datasets, standardizing features, and splitting data into training and testing sets. After training the models, performance metrics like accuracy, precision, recall, and F1-score were computed. The results showed that both models performed similarly in terms of accuracy, 82.5%, with slight differences in precision and recall. Logistic Regression excelled in precision, minimizing false positives, while Decision Trees achieved better recall, ensuring better detection of true diabetic cases. These findings highlight the importance of selecting the right model based on healthcare priorities, ultimately improving the effectiveness of preventive healthcare interventions.