Large-scale open-pit mining is a major driver of land-cover transformation in arid and semi-arid environments, yet few studies integrate spatial driver analysis with long-term predictive modeling in such landscapes. This study examines spatiotemporal land use/land cover (LULC) change at the Muruntau gold mine, Uzbekistan the world's largest open-pit gold mine operation using Landsat imagery from 1992, 2000, 2008, 2016, and 2024. LULC maps were classified with a Support Vector Machine classifier (overall accuracy 91.42%, Kappa = 0.884) across six classes: mine area, sparse vegetation, exposed rock, water, barren land, and built-up area. Five spatial drivers’ distance to the open-pit mine, roads, and water bodies, plus elevation and slope characterized the anthropogenic and topographic controls on LULC transitions and were integrated into a CA-Markov framework to simulate a business-as-usual scenario for 2032. Barren land declined by 20.0% between 1992 and 2024, accompanied by expansion of mine area (+35.5%), exposed rock (+42.0%), water bodies (+62.0%), and built-up area (+147.4%). Transition matrix analysis showed decreasing land-cover stability with decreasing barren-land persistence (0.94, 1992–2016; 0.88, 2016–2024) and increasing conversion to exposed rock. Spectral indices (SAVI, NSMI, NDTI, MBI, BSI) pointed out the southeastern sector as the main hotspot of surface disturbance. Barren land area is expected to decrease further to ~21,500 ha by 2032, while mine and exposed-rock areas continue to expand. These results offer a spatially explicit basis for land reclamation planning and environmental management in large-scale arid mining landscapes.