Abstract: Gymnadenia conopsea (L.) R.Br., commonly known as fragrant orchid, is an endangered medicinal plant species in China, predominantly distributed in the alpine and subalpine regions of Tibet, Inner Mongolia, and other ethnic minority areas. As a valuable traditional herbal medicine, it has been widely used for its therapeutic properties, making its conservation and sustainable utilization of paramount importance. Understanding the genetic diversity and phenotypic variability of G. conopsea is essential for assessing its adaptive potential, formulating effective conservation strategies, and ensuring its rational exploitation. This study systematically investigated the phenotypic and genetic diversity of G. conopsea populations across Inner Mongolia, integrating morphological trait analysis and molecular marker techniques to provide a comprehensive assessment of its genetic resources. A total of sixteen phenotypic traits were measured across six natural populations of G. conopsea in Inner Mongolia. The results demonstrated significant variability among these traits, with the coefficient of variation (CV) ranging from 15% (leaf number) to 41% (width of the sixth leaf). Among the measured traits, leaf dimensions exhibited the highest variability, suggesting strong environmental adaptability or genetic plasticity. Genetic diversity indices for phenotypic traits varied between 1.67 (leaf-related traits) and 2.07 (length of the third leaf), indicating moderate to high diversity levels. Principal component analysis (PCA) extracted five major components, collectively explaining 85.6% of the total phenotypic variation. Notably, leaf width and length were identified as the most influential factors contributing to phenotypic differentiation. Cluster analysis based on phenotypic traits classified the populations into five distinct groups: (1) broad-leaf type, characterized by wider leaves; (2) long-leaf type, exhibiting elongated leaf morphology; (3) dwarf type, with significantly reduced plant height; (4) tall type, displaying vigorous growth; and (5) multi-flower type, distinguished by a higher number of florets per inflorescence. This phenotypic classification provides valuable insights into the species' adaptive strategies across different habitats. At the molecular level, inter-simple sequence repeat (ISSR) markers were employed to assess the genetic diversity of nine G. conopsea populations comprising 125 individuals. Eight highly polymorphic ISSR primers were selected, generating a total of 89 reproducible bands, of which 82 (92.1%) were polymorphic. The genetic diversity parameters revealed considerable variation among populations: the observed number of alleles (Na) averaged 1.566, the effective number of alleles (Ne) was 1.250, Nei's gene diversity index (H) reached 0.156, and Shannon's information index (I) was 0.245. Population-level analysis indicated that the AQ and DQ populations harbored the highest genetic diversity, possibly due to their larger population sizes or more stable habitats, whereas the WU and YK populations showed relatively lower diversity, likely resulting from genetic drift or habitat fragmentation. Analysis of molecular variance (AMOVA) demonstrated that 35% of the total genetic variation existed among populations, while 65% was distributed within populations, suggesting that intra-population diversity constitutes the primary component of genetic variation in G. conopsea. The moderate genetic differentiation (Gst = 0.365) and low gene flow (Nm = 0.868) among populations imply restricted genetic exchange, possibly due to geographical isolation or ecological specialization. Bayesian clustering analysis using STRUCTURE software divided the nine populations into two major genetic clusters, with the AQ and DQ populations exhibiting admixed genetic compositions, indicating potential historical gene flow or shared ancestral polymorphism. Furthermore, association analysis between ISSR markers and phenotypic traits identified significant correlations between molecular variations and 15 morphological characteristics. The percentage of phenotypic variation explained by marker-trait associations ranged from 0.13% - 33.40%, with plant height and floret number showing the strongest associations. This study comprehensively elucidates the genetic diversity of G. conopsea in Inner Mongolia at both phenotypic and molecular levels, establishing a theoretical foundation for the conservation of genetic resources and sustainable utilization of this endangered species.