The estimation of water consumption is a crucial task in achieving global sustainability targets and addressing the long-term water needs of citizens. While some efforts have been done to estimate individual water footprints, there is still limited research in this area. To address this limitation, this article proposes a new artificial intelligence-based model, called WaterAI, to predict individuals’ water consumption scores by taking into account indirect and direct water use through the water footprint indicator. It compares four different machine learning algorithms (linear regression, LASSO regression, gradient boosting, and extreme gradient boosting) to determine the best one for water consumption estimation. The data were collected with a questionnaire survey. The experimental results show that the proposed model can be successfully used to predict personal water consumption scores in an effective way.