Water quality classification is vital for protecting universal health, preserving ecosystems, and supporting agriculture and industry. It ensures safe recreational activities, prevents economic repercussions, and promotes long-term sustainability. This study extensively analyzes machine learning-based water quality classification, focusing on key parameters like aluminum, ammonia, arsenic, barium, cadmium, using a diverse dataset with safety labels. The research preprocesses data, selects relevant features, and trains models, including Logistic Regression, K-Nearest Neighbors, SVM, Decision Trees, and various boosting algorithms. His study compared water safety using 11 machine learning models on a 70:30 training/test set. XGBClassifier, with hyperparameter tuning, achieved 97.18% accuracy. LIME and SHAP improved dataset interpretation for water quality classification.