A numerical method to simulate reactions in a cross-linked polymer is developed and applied to the photodegradation process of polyester-urethane clearcoats during artificial exposure in a Weather-Ometer. This coarse-grained simulation method, which is based on a kinetic Monte Carlo scheme, is verified with experimental data on the depth-resolved changes in optical properties and chemical composition that have been previously determined. By modelling the depth-dependency of physical processes that occur in the coating, such as the absorption of photons and the diffusion of oxygen, the experimentally observed evolution of depth gradients in chemical composition can be well described by the simulation. A sensitivity analysis of individual simulation input parameters with respect to a set of resulting observables is performed and the results provide insight into the influence of specific reaction mechanisms on the overall degradation process and help to distinguish essential from less important processes. The values of input parameters that result in the most accurate simulation of the experimental data are determined with an optimisation procedure. In this way, the numerical values of several kinetic and physical parameters that are difficult to determine directly in an experimental way, such as various reaction rate constants, can be obtained from the simulations.