This study focuses on forecasting Water Quality Index and Water Quality Classification through application of different machine learning techniques aiming to enhance predictions that are crucial for industrial operations, public health, and environmental management. Various models were applied, and their accuracy was fine-tuned using grid search optimization. Data preprocessing involved mean imputation and normalization to ensure its suitability for modeling. The dataset comprised 2005 records with seven key features. Four evaluation metrics: dataset, precision, F1-score and accuracy_score and were used to assess classification model performance. For regression models, different metrics were employed: mean squared error, median absolute error, and mean absolute error. Testing results indicated that the Gradient Boosting (GB) model achieved a WQC classification accuracy of 89.50%, producing the highest classification scores. In predicting WQI, the Multi-Layer Perceptron (MLP) regressor outperformed other models with an R2 value of 89.8%, demonstrating superior regression performance.