Teasing apart the effects of natural selection and demography on current allele frequencies is challenging, due to both processes leaving a similar molecular footprint. In particular, when attempting to identify selection in species that have undergone a recent range expansion, the increase of genetic drift at the edges of range expansions (“allele surfing”) can be a confounded factor. To address this potential issue, we first assess the long‐range colonisation history of the Aleppo pine across the Mediterranean Basin, using molecular markers. We then look for single nucleotide polymorphisms (SNPs) involved in local adaptation using: (i) environmental correlation methods (Bayenv2), focusing on bioclimatic variables important for the species’ adaptation (i.e. temperature, precipitation and water availability); and (ii) FST‐related methods (PCAdapt). In order to assess the rate of false positives caused by the allele surfing effect, these results are compared with results from simulated SNP data that mimics the species’ past range expansions and the effect of genetic drift, but with no selection. We find that the Aleppo pine shows a previously unsuspected complex genetic structure across its range, as well as evidence of selection acting on SNPs involved with the response to bioclimatic variables such as drought. This study uses an original approach to disentangle the confounding effects of drift and selection in range margin populations. It also contributes to the increased evidence that plant populations are able to adapt to new environments despite the expected accumulation of deleterious mutations that takes place during long‐range colonisations.< Réduire