When concrete is exposed to aggressive environments, it deteriorates due to reinforcement corrosion and sulfate-induced cracking, both of which are caused by chloride ions and sulfates, respectively. Evaluating these challenges using conventional laboratory methods is often labour-intensive and may fail to capture the complex nonlinear relationships between influential parameters such as mix design, curing conditions, and exposure duration. This study proposes a data-driven alternative by applying machine learning (ML) algorithms to predict two major durability indicators: chloride ion penetration and sulfate resistance. Four supervised ML models: Linear Regression, Decision Tree, Random Forest and XGBoost, were developed and trained on two independent datasets comprising over 1750 data points. Key preprocessing steps used on the datasets include handling missing values, feature selection, and normalization. The models’ performances were evaluated using standard metrics: Mean Absolute Error (MAE), Root Mean Squared Error (RMSE) and Coefficient of Determination (R2). The Random Forest model achieved the highest performance in predicting chloride ion penetration, with an R2 of 0.73, indicating its strong capability to model complex interactions. For sulfate resistance, XGBoost delivered the most accurate results with an R2 of 0.95, demonstrating its robustness and generalization ability. Feature importance analysis further revealed the water-binder ratio and cement content as influential predictors across both durability mechanisms. The results underscore the potential of ML as a reliable, efficient, and scalable tool for durability assessment. This approach can enhance decision-making in concrete mix design, reduce reliance on physical testing, and contribute to the development of more resilient infrastructure systems.