<p>Water pollution presents considerable challenges to ecosystems, human health, and economic stability, underscoring the importance of robust monitoring and management systems. Water quality monitoring and management remain critical global challenges, particularly in achieving UN Sustainable Development Goal (SDG) 6. This research leverages Machine Learning (ML) techniques to predict the Water Quality Index (WQI) for Haridwar, a region in North India, utilising a comprehensive dataset that includes parameters like pH, Dissolved Oxygen (DO), Biochemical Oxygen Demand (BOD), and ion concentrations. Multiple ML models were employed, including Support Vector Regression (SVR), Random Forest Regressor (RFR), Extreme Gradient Boosting Regressor (XGBR), and Recurrent Neural Network (RNN). The models were evaluated using various performance metrics; the RNN model demonstrated superior performance, achieving an RMSE of 0.7535, MAE of 0.5742, MSE of 0.3004, and an R² score of 0.9655. The WQI for the various sites considered ranged between 90 and 110, with an average WQI of 96 indicating moderate water quality conditions, revealing seasonal fluctuations, and periods of moderate pollution indicating potential water quality concerns. These results emphasize the effectiveness and precision of ML-based methods for water quality prediction, providing a valuable tool for proactive environmental management. This study highlights the potential of such models in promoting sustainable river ecosystem management, aligning with the objectives of SDGs 6 and 11. Furthermore, the methodological approach is versatile and can be adapted for water quality assessments in various geographical settings.</p>