The textile industry is essential to the global economy but poses environmental challenges, particularly regarding water use and pollution. Sustainable production necessitates improved water management and reuse. This study forecasts water quality for reuse using data collected from twelve Portuguese textile companies over two months. To achieve this, the CRISP-DM methodology was adhered to, monitoring key water quality factors such as pH and turbidity. The data were cleaned, features selected, normalized, and class imbalance addressed using SMOTE. Five machine learning models were assessed: Decision Trees, Random Forest, Logistic Regression, Support Vector Machines, and XGBoost. The best results were obtained from XGBoost, Random Forest, and Logistic Regression, particularly in scenario 2. In Company 11, Random Forest achieved 98% sensitivity and a 97% F1-score, whereas XGBoost exhibited 98% sensitivity and 94% accuracy. These models demonstrated high accuracy and sensitivity in classifying water quality. The findings indicate that machine learning can assist in enhancing water reuse decisions in the textile industry and mitigating environmental impact.