Existing water quality index (WQI) models assess water quality using a range of classification schemes. Consequently, different methods provide a number of interpretations for the same water properties that contribute to a considerable amount of uncertainty in the correct classification of water quality. The aims of this study were to evaluate the performance of the water quality index (WQI) model in order to classify coastal water quality correctly using a completely new classification scheme. Cork Harbour water quality data was used in this study, which was collected by Ireland's environmental protection agency (EPA). In the present study, four machine-learning classifier algorithms, including support vector machines (SVM), Naïve Bayes (NB), random forest (RF), k-nearest neighbour (KNN), and gradient boosting (XGBoost), were utilized to identify the best classifier for predicting water quality classes using widely used seven WQI models, whereas three models are completely new and recently proposed by the authors. The KNN (100% correct and 0% wrong) and XGBoost (99.9% correct and 0.1% wrong) algorithms were outperformed in predicting the water quality accurately for seven WQI models. The model validation results indicate that the XGBoost classifier outperformed, including accuracy (1.0), precision (0.99), sensitivity (0.99), specificity (1.0), and F1 (0.99) score, in order to predict the correct classification of water quality. Moreover, compared to WQI models, higher prediction accuracy, precision, sensitivity, specificity, and F1 score were found for the weighted quadratic mean (WQM) and unweighted root mean square (RMS) WQI models, respectively, for each class. The findings of this study showed that the WQM and RMS models could be effective and reliable for assessing coastal water quality in terms of correct classification. Therefore, this study could be helpful in providing accurate water quality information to researchers, policymakers, and water research personnel for monitoring using the WQI model more effectively.