The extensive use of predictive models in healthcare raises significant challenges in dealing with patient data privacy and adherence to laws like the Health Insurance Portability and Accountability Act (HIPAA) and the General Data Protection Regulation (GDPR). Federated Learning (FL) presents a viable alternative by enabling cooperative model training over dispersed datasets, protecting sensitive patient data security and confidentiality. This study explores the use of hybrid, vertical, and horizontal FL paradigms in predicting heart disease. The study concentrates on how these strategies handle issues with data distribution, increase privacy by utilizing methods like homomorphic encryption and differential privacy, and eventually raise predictive model accuracy. This study highlights FL’s ability to transform healthcare by analyzing real-world use cases and metrics. It sets new benchmarks for ethical artificial intelligence (AI) in medicine by illustrating how privacy-preserving machine learning may produce precise illness prediction while protecting patient confidentiality.