The durability of aggregates is an important factor that is used as an input parameter in desirable engineering properties along with resistance to exposure conditions. However, it is sometimes difficult to determine the durability of aggregates in the laboratory (with a magnesium sulfate test) because this test is time-consuming and expensive. In this paper, the physical and mechanical properties including water absorption and the Los Angeles coefficient are tailored to the specific evaluation of the durability of limestone aggregates. However, the predictive capabilities of artificial neural networks (ANN) and hybrid particle swarm optimization-based (PSO) ANN techniques have been evaluated and compared using the same input variables. To assess the reliability of the model, some performance indices such as the correlation coefficient (R (2)), the variance account for, and the root-mean-square error were calculated and compared for the two developed models. The results revealed that even though the two developed models reliably predict the durability value (magnesium sulfate value), the proposed PSO-ANN method displays an obvious potential for the reliable assessment of the value of magnesium sulfate according to the model performance criterion.