Semi-arid mountain protected areas are increasingly exposed to interacting climatic and human-induced hazards that threaten ecosystem stability and conservation planning. This study develops a high-resolution, spatially explicit expert-analytical framework to assess multi-hazard exposure in the Sheida Protected Area (23,832 ha) in Iran's Zagros range. Eight major hazards-drought, soil erosion, evapotranspiration, high-wind events, wildfires, extreme temperatures, floods, and landslides-were quantified using long-term meteorological records and high-resolution geospatial datasets. A Delphi-based expert elicitation (n = 10; Cronbach's alpha = 0.94) was used to derive hazard weights, which were integrated with standardized spatial indicators across a 1-km2 grid (n = 1122 cells) to construct a composite exposure index. Spatial autocorrelation analyses (Global Moran's I and LISA) identified strong clustering of exposure intensities, with 18.3% of the landscape forming significant High-High hotspots in the southern sector. Drought (weight = 0.181) and wildfires (0.168) emerged as dominant drivers, consistent with regional climate-driven stress patterns. A support vector machine (SVM) model (AUC = 0.90 training; 0.85 validation) further indicated high wildfire susceptibility across 65% of the area. The results provide a robust, scalable approach for multi-hazard assessment in semi-arid mountainous social-ecological systems and delineate priority zones for fire management, soil conservation, watershed stabilization, and climate-adaptive ecosystem stewardship.