Ecological restoration programmes aiming at population recovery of imperilled plant species increasingly involve plant translocations. Evaluating the genetic status of seed source and target populations is essential for designing plant translocation protocols and optimizing recovery success. We developed nine polymorphic microsatellite markers and used three plastid markers to investigate genetic variation and structure of the two last large and six small remaining populations of the self-incompatible, clonally-propagating Arnica montana in southern Belgium and bordering France. The aim of the study was to determine the genetic status of these remaining populations and whether the large populations can be used as seed source for translocations. Most small populations maintained high genetic diversity and showed no inbreeding or a heterozygote excess, which may be explained by high genet longevity thanks to clonal propagation, heterosis, inbreeding depression at early development stages and/or no recruitment. Genotypic diversity was low in small populations, with clonal propagation mainly contributing to rosette production. The number of genets, and therefore effective population size, was often very small, restricting compatible mate availability. The situation is therefore more critical than it seems on the field, and bringing new genetic variation is necessary. Although no polymorphism was found in plastid DNA markers, between-population differentiation based on microsatellite markers was moderate, except for very small populations, where it was greater ( F ST  > 0.200). These patterns of differentiation were likely due to genetic drift effects and demographic stochasticity. We recommend using mixed seed material from the two large populations for translocations, and before conducting reinforcements, to first implement crossing experiments and reintroductions of mixed and crossed material in ecologically restored sites to understand the long-term effects of combining genotypes from different locations.