Protecting both the environment and public health depends on access to clean water. This study evaluates five advanced ML models to assess their effectiveness in estimating water quality (WQ). The models examined include Light Gradient Boosting Machine (LGBM), CatBoost, AdaBoost, XGBoost Classifier (XGBC), and XGBoost combined with Ant Colony Optimization (XGBC + ACO). The hybrid XGBC + ACO model proved superior, effectively capturing complex patterns in water quality datasets. While standalone XGBC and AdaBoost generated strong predictions, the hybrid model was robust enough to be recognized as the best. Optimized machine learning models can accurately predict water quality, enabling informed decisions for effective water resource management and environmental conservation. XGBC + ACO has shown better results, an accuracy of 96.8% and an F1 score of 0.98, emphasizing the value of integrating optimization techniques with machine learning algorithms.