Accurate brain tumor segmentation is vital for effective diagnosis and treatment planning. Manual segmentation methods, though precise, are labor-intensive and susceptible to variability. Automated segmentation using deep learning, particularly Convolutional Neural Networks (CNNs), has shown promise in enhancing accuracy and consistency. However, the adoption of such technologies in healthcare faces significant hurdles due to concerns over data privacy and the need for data sharing across institutions. This study addresses these challenges by implementing Federated Learning (FL), a model that facilitates decentralized training without direct data exchange, thereby preserving data privacy. Using the BraTS2020 dataset, we demonstrate that a Hybrid U-Net model, enhanced with one-hot encoding, effectively segments brain tumors while maintaining patient data confidentiality. Our results confirm FL’s effectiveness, achieving a Dice Score of 0.9826, thus confirming the potential of FL in sensitive healthcare applications.