Polymer-modified asphalt (PMA) mixtures offer improved performance compared with conventional binders, particularly in terms of crack resistance and durability under increasing traffic and environmental loading. However, capturing their micromechanical behavior and failure mechanisms remains challenging owing to their internal heterogeneity and particle-scale interactions. This study utilized the Discrete Element Method (DEM), implemented via the Particle Flow Code (PFC), to simulate Uniaxial Compressive Strength (UCS) and Indirect Tensile Strength (ITS) tests on PMA mixtures with varying specimen sizes and aggregate gradations. Two contact models, Linear Contact Bond (LCB) and Soft Bond (SB), were evaluated to represent tensile fracture and progressive bond degradation. The numerical models were calibrated and validated against the experimental results, which showed a deviation of less than 1.5 % in the compressive strength and modulus across the nine PMA formulations. A strength-size effect was observed and modeled, enabling the conversion of non-standard field core strengths to laboratory-equivalent values. Additionally, a high-resolution image dataset of asphalt surface distress (cracks, potholes, and marking degradation) was developed to support computer vision–based pavement monitoring. The integrated simulation and imaging framework presented in this study offers new insights into microscale failure behavior, supports more accurate field data interpretation, and contributes to intelligent maintenance strategies for resilient pavement infrastructure.