Aim To assess how climate change will affect the distribution of rupestrian plant species in the Iron Quadrangle (IQ), southeastern Brazil, and to identify conservation priorities under future scenarios. Location Iron Quadrangle, Minas Gerais, Brazil. Taxon Eight rupestrian angiosperms: Aiouea tetragona (Lauraceae), Dyckia consimilis, D. rariflora, Hoplocryptanthus ferrarius, H. schwackeanus, Vriesea minarum (Bromeliaceae), Paepalanthus amoenus and P. magalhaesii (Eriocaulaceae). Methods We compiled occurrence records from online herbaria and Vale S.A. datasets, applied spatial thinning, and fitted ensemble ecological niche models (ENMTML). Current suitability was estimated from WorldClim v2.1 bioclimatic variables; future projections used three CMIP6 general circulation models under SSP126 and SSP585 for 2050 and 2090. We overlaid suitability with protected areas to quantify overlap, identify candidate climatic refugia, and map potential corridors. Results All species achieved acceptable to excellent predictive performance (AUC > 0.75; TSS > 0.5). Models project widespread contractions of suitable climate space, with xeromorphic taxa ( A. tetragona, D. rariflora, H. ferrarius, P. magalhaesii) approaching zero suitability by 2090 under SSP585. D. consimilis and H. schwackeanus were comparatively stable, whereas P. amoenus and V. minarum showed transient gains under intermediate scenarios. Persistent hotspots were concentrated along the Gandarela–Caraça axis (e.g., Serra do Gandarela and the RPPN Horto Alegria), although dispersal is constrained by edaphic specialization. Main conclusions Mitigation alone will not safeguard rupestrian flora in the Iron Quadrangle. Proactive actions—strengthening ecological connectivity, establishing seed banks and germplasm collections, and planning assisted translocations—are urgently needed. We provide an interactive tool that translates model outputs into management-ready products to prioritize interventions in landscapes such as the Gandarela–Caraça corridor.