Understanding adaptive evolution and survival risks in understory herbs is crucial for the effective conservation of biodiversity. It is not well understood how environmental gradients mediate local adaptation and how populations respond to changing environment of understory herbs. In this study, we conducted the population genomic analysis of Adenocaulon himalaicum (Asteraceae) in East Asia, representing a good model of dominant understory herb to unveil the adaptive evolution mechanisms of this species. Based on 34,398 single nucleotide polymorphisms (SNPs) across 27 populations, we identified three genetic lineages companied with high levels genetic differentiation between populations. Our isolation by environment results (IBE) indicated the significant effect of environmental gradients on genomic variation of A. himalaicum (r = 0.18, p = 0.03). Redundancy analysis (RDA) revealed that the genomic variation were significantly explained by mean diurnal range (bio2), precipitation of driest quarter (bio17) and precipitation of warmest quarter (bio18) and soil type (TAXNWRB). Using two genotype-environment association (GEA) methods, we identified 27 SNPs as candidates for core climate-related adaption loci and two loci of these were further validated by qRT-RCR experiments. Furthermore, predictions to spatiotemporal pattern of genomic offsets under different future climate scenarios revealed that populations near southeastern edge of Himalaya, Yunnan-Guizhou Plateau, northern Korean Peninsula, and southern Japan were regarded as the most vulnerable populations. Therefore, conservation priority is needed for these populations due to candidate variants and unique genetic compositions being useful for future breeding.