ABSTRACT This study investigates the effects of ground tire rubber (GTR) on the cracking resistance of asphalt mortar from different analysis scales. Furthermore, the effectiveness of different fracture test methods was compared for GTR-modified asphalt mortar, including the indirect tensile cracking test (IDEAL-CT), disk-shaped compact tension test (DCT) and Illinois flexibility index test (I-FIT). To achieve this, a mesoscopic fracture test based on the SEM-Servo Plus device was proposed firstly to obtain the micro-parameters of constitutive models in discrete-element modelling (DEM). A unified fracture modelling procedure that can avoid any interference factors was carried out for comparative fracture analysis. Results from the tests and simulations show that rubber filler plays a positive role in enhancing the fracture properties of the asphalt mortar. The fracture energy measured by IDEAL-CT is almost 5–6 times of the DC(T) test and 1.5–2 times of the I-FIT test. It is not an inherent attribute of material and can only be regarded as a relative index for fracture resistance evaluation. The FI and CTindex obtained from the I-FIT and IDEAL-CT test show a weak correlation of fracture resistance, especially when comparatively evaluating the GTR-modified asphalt mortar with minor difference in filler content.