Bituminous mixtures in pavements can experience fatigue due to the repeated traffic loading, which leads to cracking of the bituminous layers. In view of the significance of fatigue cracking, extensive fatigue testing has been carried out on both bituminous binders and bituminous mixtures. This research focuses on evaluating and improving the existing test methods for quantifying fatigue resistance of the material. First, in tests on bituminous binders such as the Time Sweep and the Linear Amplitude Sweep tests, the influence of secondary flows and flow instabilities on the decrease in dynamic shear modulus is determined by conducting oscillatory tests at different gaps. The lower the separation between the parallel plates, the higher was the fatigue resistance predicted by these test methods. This behaviour is similar to that observed for polymer solutions and polymer melts by researchers, and can be attributed to the instabilities associated with torsional flows of nonlinear viscoelastic materials. Following this, a new methodology was developed for analyzing the results of four-point beam (4PB) fatigue tests on bituminous mixtures. In the 4PB fatigue test, the strain varies along the length and the depth of the beam due to which the modulus of different parts of the beam decreases at different rates. In traditional methods of analysis, the modulus is assumed to decrease uniformly at all parts of the beam, and thus, the calculated modulus is essentially a weighted average of the modulus at different locations. In the newly proposed methodology, the evolution of local modulus corresponding to the said applied strain level in 4PB tests can be computed. A new fatigue failure criterion is proposed based on the evolution of local modulus. Furthermore, harmonic analysis was used to characterize the fatigue behaviour of bituminous mixtures in strain-controlled 4PB tests. The stress response from each loading cycle was split into different harmonics using the Fast Fourier transform. The ratio of the higher harmonics to the fundamental harmonic of the stress response was found to sharply increase after a certain number of cycles. Considering this, the number of cycles corresponding to the increase in the stress amplitude ratio was identified as the failure point of the beam in these fatigue tests. This criterion was in agreement with the local modulus criterion mentioned earlier. Furthermore, the energy dissipated in repeated loading due to the initiation and propagation of cracks, also referred to as the work of fracture, was evaluated by conducting uniaxial pull-off tests on bitumen sandwiched between two rock disks. By conducting tests at different loading rates, the proportion of adhesive-and of cohesive failure was determined. After determining the total dissipation using the load-displacement data, the dissipation due to viscous friction was estimated using a viscoelastic model, and the work of fracture was then determined by subtracting the viscous dissipation from the total dissipation. The work of cohesion of the bituminous mastic and the work of adhesion between bituminous mastic and aggregate at the interface were then estimated using the work of fracture and the proportion of cohesive and adhesive failure at each loading rate. The findings show that the higher the work of cohesion, the longer the fatigue life of bituminous mixtures.