<p>As global demand rises and natural aggregates become scarcer, using supplementary materials and industrial waste for concrete production offers a sustainable solution that conserves resources and reduces waste. Previous studies have demonstrated that pozzolanic additives, such as natural pozzolans and ceramic waste powders, can enhance concrete performance by refining the pore structure, reducing permeability, and improving long-term strength. However, limited research has explored the synergistic use of these materials, particularly in the context of predictive modeling and performance optimization. In the current study, an investigation was conducted to evaluate the effects of replacing cement with waste ceramic powder (at 10%, 20%, and 30%) and natural pozzolan (at 5%, 10%, 15%, and 20%) on the performance of concrete. The physical, mechanical, and durability properties of the resulting concrete mixes were comprehensively assessed, including compressive, tensile, and flexural strength, water absorption, and permeability. These properties were evaluated at various curing ages —7, 28, and 91 days—for compressive, tensile, and flexural strengths and water absorption, while permeability was measured at 91 days. Machine learning models—namely XGBoost, Random Forest, Gradient Boosting, and Elastic Net, were employed to predict compressive, tensile, and flexural strengths, with XGBoost yielding the highest accuracy for compressive and tensile strengths (RMSE = 0.8 and 0.08, respectively) and Gradient Boosting excelling for flexural strength (RMSE = 0.2). The results identified the Kh15C10 mix design (15% Natural pozzolan and 10% waste ceramic powder) as the optimal configuration, exhibiting enhanced compressive strength (8.12–11.49% increase), tensile strength (3.60–6.33% increase), flexural strength (2.99–12.45% increase), as well as reduced water absorption (35.74–77.33% decrease) and permeability (99.82% decrease) compared to the control sample. The integration of waste ceramic powder and natural pozzolan not only enhances mechanical and durability performance but also significantly contributes to environmental sustainability by lowering cement demand, reducing CO₂ emissions, and promoting circular economy principles through waste valorization. In predictive modeling, XGBoost excelled for compressive and tensile strengths with the lowest RMSE values (0.8 and 0.08, respectively), while Gradient Boosting outperformed for flexural strength with an RMSE of 0.2. These findings highlight the potential of these models in optimizing sustainable concrete designs.</p>