Background <p>Understanding how ecotypic divergence persists under extensive gene flow is critical for predicting adaptive responses in long-lived conifers. Mongolian Scots pine (<i>Pinus sylvestris</i> var. <i>mongolica</i>) occupies contrasting mountain and sandy-dune habitats in northeastern China, forming two ecotypes with distinct environmental adaptations. Using reference-free Specific Locus Amplified Fragment sequencing (SLAF-seq) data and genome-wide SNPs, we explored patterns of genomic differentiation and genotype-environment associations to determine whether adaptive divergence is driven by few large-effect loci or polygenic shifts.</p> Results <p>Despite weak population structure, we identified ~ 3% of the genome as adaptive outliers with strong differentiation between ecotypes. These loci exhibited significant environmental associations indicating that both adaptive genomic islands and polygenic shifts are dominated by large-effect variants rather than minor-effect alleles. Furthermore, phenotypic differentiation between ecotypes reflected a dual mechanism: adaptive genetic divergence shaping hydraulic traits and mechanical support, and genotype-by-environment (G×E) interactions enabling phenotypic plasticity in physiological responses, such as water-use efficiency and stress tolerance.</p> Conclusion <p>The interplay between hard genetic adaptation and plasticity highlights how Mongolian Scots pine can simultaneously preserve ecotypic differentiation and respond flexibly to environmental heterogeneity. Importantly, such plasticity may provide a critical buffer against rapid climate change, allowing populations to persist despite ongoing gene flow and delayed genomic divergence. Our study highlights how divergent selection and plasticity together maintain ecotypic divergence in wind-pollinated pine variety with high gene flow and demonstrates a novel reference-free approach for dissecting adaptive architectures in non-model genomes.</p>