Demand prediction is a significant component of managing supply chains because it can increase profit and efficiency by identifying supply channels with predicted consumer demand. The data appropriately chosen to enhance corporate sales and profits and also to identify the customer buying patterns and behavior. This paper proposes an integrated data-driven cluster-based demand forecasting approach for successful retail businesses, segmenting customers based on Recency, Frequency, and Monetary (RFM). After the segmentation of customers, the prediction of demand for each segment is carried out by Long Short-Term Memory (LSTM). The results show that consumer segmented prediction gives better accuracy of MAPE = 8.23, MAE = 17.45, RMSE = 145.8. Since accurate and precise demand forecasting is crucial for efficient inventory and supply chain management, the current study tackles one of the most severe problems in retail management. The suggested model performs better than other popular existing demand forecasting techniques.