This paper presents a real-time prediction system for river discharge using clustering techniques in data mining, specifically focusing on the Brahmaputra River's discharge data from 2003 to 2007. By employing agglomerative hierarchical clustering and k-means clustering, the study predicts discharge patterns for the subsequent year, which aids in effective flood forecasting. The methodology includes statistical analysis, data standardization, and validation using the coefficient of determination to evaluate the accuracy of the predictions.