This paper proposes a comprehensive methodology for improving supply chain management through Walmart’s Electronic Data Interchange (EDI) system. The approach begins with the acquisition of standardized supply chain data from Walmart’s EDI, encompassing key elements such as inventory, orders, and shipping details. Given the presence of missing data in such datasets, a K-Nearest Neighbors (KNN)-based imputation technique is applied to fill in the gaps by leveraging patterns in neighboring data points. To enhance the accuracy of the subsequent analysis, Fisher score-based feature selection is used to retain the most impactful features affecting supply chain performance. Lastly, a Long Short-Term Memory (LSTM) neural network is employed to forecast demand trends, optimizing inventory management through accurate time-series predictions. This integrated methodology aims to improve decision-making, operational efficiency, and forecasting accuracy within Walmart’s supply chain ecosystem.