The optimization of road lighting installations is nowadays mainly considered from the point of view of reducing energy costs by using the advantages of LED technology when designing or renovating lighting installations. Knowing the actual reflection properties of the road to be illuminated enables further optimization but these are too often not measured. Moreover, the evolution of these properties according to the surface state of the road (dry, moist, wet, soaked) is never considered. This paper proposes a methodology to design an adaptive lighting system that takes into account the optical properties of the pavement according to the weather conditions in order to optimize the visibility offered to users. This methodology is based on numerous simulations of potential pavement surface states. By comparing the developed adaptive lighting system with a conventional lighting system, it is shown that the overall optimization of the lighting operation is achieved both on photometric and on energy aspects.