This paper presents a computational modelling approach to evaluate moisture damage in hot mix asphalt (HMA) materials. The approach combines a methodology to randomly generate HMA microstructures and a finite element (FE) formulation that uses a coupled mechanical-moisture continuum damage constitutive relationship. Probable two-dimensional HMA microstructures were randomly generated and treated as independent replicates. These replicates were used as the model geometry in FEs and they were computationally subjected to a one-year moisture-conditioning regime. During this time, a mechanical load was periodically applied. The mean and dispersion values related to the evolution of moisture and mechanical damage were identified, and the effects of varying the air void content and the moisture diffusion coefficient of the asphalt matrix were assessed. The results demonstrate that the modelling approach is a viable tool to probabilistically quantify the impact of physical and volumetric properties of HMA mixtures on their susceptibility to undergo moisture-mechanical damage.