This research addresses the escalating prevalence of chronic diseases and the associated rise in mortality rates. Early detection of these conditions is paramount for improved patient outcomes. We present a novel artificial intelligence model for the prediction of myocardial infarction, leveraging a neural network architecture enhanced by particle swarm optimization. This optimization technique facilitates the identification of salient features, thereby maximizing predictive accuracy. The proposed model achieved an accuracy of 90%, demonstrating the critical influence of input data quality. Comparative analysis against established machine learning methodologies, including Random Forest, Deep Learning, and Support Vector Machines, revealed superior performance and computational efficiency. The results suggest the potential for this model to be implemented in clinical settings for rapid and accurate diagnosis, and for the development of accessible, patient-facing health monitoring tools.