AbstractThis study used an advanced modelling approach capable of capturing the complex behaviour of asphalt concrete to model the modified wheel tracking test using a recent advanced experimental test set-up in accordance with ASTM D8292-20. The modelling approach uses the discrete element method (DEM) to naturally produce the heterogeneous internal structure and governs the behaviour of asphalt concrete at the grain level by an interparticle contact model. The contact model used is capable of characterising the rate and time dependency, viscoelastic-damage, and plastic-damage behaviour of asphalt concrete utilising the coupling of an elastoplastic-damage law with a viscoelastic-damage law. Unlike the conventional wheel tracking tests run in a fixed boundary condition (fully confined), the modified wheel tracking test considers the effect of boundary conditions on the rutting behaviour of asphalt mixes. Through comparisons and verifications with laboratory data of the rutting test at different boundary conditions (fully confined and unconfined), the modelling approach shows its capability of capturing the rutting behaviour of asphalt concrete in the modified wheel tracking test. Micromechanics analysis shows that the third (tertiary) stage of rutting behaviour is due to the weakening of the internal structure of the asphalt samples with contact bond breaks over time, which is found in the unconfined test. Meanwhile, the tertiary stage hardly occurs in the fully confined test once densification leads to contact of the aggregate–aggregate skeleton, forming a rigid structure to resist the load with lateral support from the fixed boundary condition. Finally, a parametric study was also conducted to provide further insight into the current testing set-up, including the effect of the sample size and boundary condition on the rutting behaviour of asphalt concrete.