<p>Evaluating water quality is essential for environmental sustainability, public health, and ecosystem protection. Nowadays, the evaluation of river water quality is a critical issue for domestic demand, industrial use, aquatic life, and agricultural production throughout the world. The basic objective of this investigation was to forecast the WQI, or water quality index, of the Gumara and Ribb rivers by using various machine learning models. The WQI of these rivers was trained and tested using eight ML algorithms and 22 water quality parameters in Python. These machine learning models are Support Vector, Random Forest, Decision Tree, Logistic Regression, k-Nearest Neighbors, Gradient Boosting, Naive Bayes, and AdaBoost. The total dataset used for this study consisted of 4598 samples. The statistical ML models’ performance was assessed using RMSE, MARE, NSE, and R². The results of this study showed that the highest values of R², MARE, RMSE, and NSE during the training and testing of models were 0.998, 4.6, 5.2, and 0.98, and 0.978, 3.23, 4.15, and 0.965, respectively. The class-wise ML models’ performance was assessed using F1 score, precision, accuracy, recall, and AUC. The highest results of class-wise model performance evaluation for accuracy, F1score, precision, and recall were 0.999, 0.985, 0.978, and 0.975, respectively. The highest value of the ROC-AUC curve in this study was 0.93 during GB and RF; this indicates that these models were the best for forecasting the WQI of this study. Feature importance is typically used to select the most influential parameters of water quality; for this study, NO₃ was the most significant factor among the water quality parameters. Therefore, to accurately assess and predict the WQI of river water, using advanced ML models is most significant for protecting against water pollution and environmental degradation.</p>