This research is focused on damage development in the adhesive zone and the mortar of porous asphalt concrete. The motive of this research is the loss of stone from the pavement surface, the so-called ravelling of noise reducing surface wearing courses. Ravelling is the dominant defect of porous asphalt resulting in huge costs of maintenance and resurfacing in the Netherlands. Ravelling is a mixture-associated problem and is directly related to the binding failure within the stone-to-stone contact regions. This research is thus towards a better understanding of the processes responsible for ravelling, i.e. cohesive failure in the mortar bridge and adhesive failure at the mortar/aggregate adhesive zone on the basis of meso-mechanics. An intensive experimental program was carried out on various adhesive zones and mortars at a meso-sale of millimetres. This program was designed based on a better understanding of stress/strain states in porous asphalt concrete under moving traffic loadings. The aim was to develop fatigue/damage models for the adhesive zone and mortar which allows life expectancy to be predicted. Tension and shear tests as well as tests in which tension and shear were combined were performed on the adhesive zones. Among these tests, uniaxial tension testing was conducted using Dynamic Mechanics Analyzer. Shear testing was performed using Dynamic Shear Rheometer. A test which combined tension and shear was specially designed. Various types of load signals were applied accounting for complex stress signals to which the adhesive zones are subjected in practice. A damage model based on a linear cumulative damage rule was used to explain the obtained test data. The proposed model is making use of the integration of an equivalent uniaxial tensile stress signal to compute the development of damage. To do this, an internal-friction theory was applied to translate any combination of shear and normal stresses into a simple equivalent uniaxial tensile stress. Model fit indicated that the predicted number of cycles to failure is in agreement with the measured data. A practical mortar fatigue model based on the dissipated energy concept was developed for the life predictions under complex multiaxial loadings. The fatigue model is based on the initial dissipated energy per cycle. Model parameters can be determined on the basis of commonly used fatigue tests. The proposed model gave very good results in explaining the laboratory fatigue data. By combining the mortar stress and strain signals provided by response calculations the dissipated energy per tyre passage can be calculated and the number of load repetitions to failure can be estimated using the mortar fatigue model. The developed fatigue/damage models were applied to explain the ravelling damage of four different porous asphalt mixtures used in a full-scale accelerated pavement testing (APT) experiment. Finite element simulations of the behaviour of these four mixtures were made. A 2D idealized model was used for practical reasons. The life expectancy predicted by this way was in good agreement with the results of full scale ravelling tests. The life time predictions as made explain the initiation of ravelling. In reality however maintenance will only be applied if ravelling has extended over a certain area and at a certain level of severity. To relate the computed life expectancy with the actual lifespan of porous asphalt or the maintenance moment, a ravelling development model was proposed. By combining this model with the initial ravelling damage obtained from the simulations, the moment for maintenance and resurfacing can be estimated for planning pavement preservation strategies. Finally, it was concluded that the meso-scale fatigue characterization of the adhesive zone and mortar is feasible. The developed adhesive zone damage model and mortar fatigue model together with the finite element model for the prediction of stress, strain and dissipated energy allow to arrive at a proper ranking in performance of porous asphalt mixtures. It has been shown that the developed models are very useful tools to design porous asphalt concrete with a better raveling resistance and a longer lifetime.