Water quality prediction is a critical task with implications for public health and environmental management. In this study, we introduce a novel ensemble approach, termed the voting classifier, to predict water quality. Our approach integrates diverse machine learning models, including neural networks, k-nearest neighbor (k-NN), random forest, and decision tree. The dataset is carefully pre-processed (missing values are handled, text representations are converted to numeric values, and feature normalization is accomplished using min–max scaling). The dataset is derived from water quality measurements. The use of a neural network, a potent machine learning technique renowned for its intricacy and versatility, is a unique aspect of this work. After dividing the dataset into training and testing sets (80–20 split), the scaled data is used to train the individual random forest, decision tree, k-NN, and neural network models. The neural network achieved 91% accuracy, according to the data. The combined accuracy of the models when they are incorporated into the voting classifier ensemble with the other models is an astounding 95%. This significant improvement in accuracy emphasizes the power of integrating various models in an ensemble, highlighting the ability of various methods to work well together. In addition to showing the effectiveness of ensemble learning in predicting water quality, the study emphasizes the significant influence of integrating a neural network into the ensemble. The results provide important new information to the field of water quality prediction by demonstrating how adding a neural network to a combination of machine learning models can greatly enhance predictive performance.