This paper explores the integration of Federated Learning (FL) and Edge Computing (EC) to optimize supply chain operations in a decentralized and privacy-preserving manner. Outdated supply chain management methods face the challenges like data silos, slow decision-making, and limited real-time visibility. The proposed FL–EC framework addresses these issues with individual nodes warehouses, transport fleets, and retailers to train machine learning copies minus sharing the sensitive statistics collaboratively. This improves data privacy, allowing real-time optimization in areas with demand forecasting, inventory management, and route planning. Edge computing improves the efficiency in processing data, reducing latency and faster responses to disruptions. The framework is evaluated against the traditional methods, demonstrating superior operational efficiency, cost reduction, and data privacy. The study identifies the key challenges, including data heterogeneity, communication delays, and privacy concerns, and suggests future directions for improving the scalability, model aggregation, and security. This work contributes a novel approach to supply chain management, which is intelligent and sustainable, paving the way for resilient global supply chains.