<p>The mix design of cement-based concrete composites involves considering multiple parameters to achieve a desired compressive strength, often requiring extensive laboratory experimentation to optimize different mixture proportions, particularly when incorporating admixtures. Hence, leveraging modern techniques such as machine learning becomes essential to address this challenge, facilitating the optimization of time, energy, and cost-effectiveness. This study employs six different models namely linear regression model (LRM), non-linear regression model (NLR), artificial neural network (ANN), gaussian process regression (GPR), support vector machine model (SVM), and ensemble regression model (ER) to develop predictive models aimed at forecasting the compressive strength of concrete enhanced with Arabic gum. The performance of the developed models in predicting compressive strength for concrete mixtures was evaluated using four different statistical parameters, including the coefficient of determination (R²), Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), and scatter index (SI). The analysis revealed that, among the models evaluated, the Artificial Neural Network model significantly outperforms the others. It demonstrated superior performance with an R² value of 0.99, indicating a high level of accuracy, and lower error metrics, including RMSE (0.38), MAE (0.289), and SI (0.011), when compared to the other developed models. These results highlight the ANN model’s robustness and effectiveness in predicting the compressive strength of Arabic Gum-modified concrete, making it the most reliable option among the tested models.</p> Graphical abstract <p></p>