Manipulating evolutionary forces imposed by hosts on pathogens like genetic drift and selection could avoid the emergence of virulent pathogens. For instance, increasing genetic drift could decrease the risk of pathogen adaptation through the random fixation of deleterious mutations or the elimination of favorable ones in the pathogen population. However, no experimental proof of this approach is available for a plant-pathogen system. We studied the impact of pepper ( Capsicum annuum ) lines carrying the same major resistance gene but contrasted genetic backgrounds on the evolution of Potato virus Y (PVY). The pepper lines were chosen for the contrasted levels of genetic drift (inversely related to N e , the effective population size) they exert on PVY populations, as well as for their contrasted resistance efficiency (inversely related to the initial replicative fitness, W i , of PVY in these lines). Experimental evolution was performed by serially passaging 64 PVY populations every month on six contrasted pepper lines during seven months. These lines exhibited highly divergent evolutionary trajectories, ranging from viral extinctions to replicative fitness gains. The sequencing of the PVY VPg cistron, where adaptive mutations are likely to occur, allowed linking these replicative fitness gains to parallel adaptive nonsynonymous mutations. Evolutionary trajectories were well explained by the genetic drift imposed by the host. More specifically, N e , W i and their synergistic interaction played a major role in the fate of PVY populations. When N e was low ( i . e . strong genetic drift), the final PVY replicative fitness remained close to the initial replicative fitness, whereas when N e was high ( i . e . low genetic drift), the final PVY replicative fitness was high independently of the replicative fitness of the initially inoculated virus. We show that combining a high resistance efficiency (low W i ) and a strong genetic drift (low N e ) is the best solution to increase resistance durability, that is, to avoid virus adaptation on the long term.