Extreme events (e.g., drought and heatwave) occur frequently and intensively with climate change, where the combination of dry and hot events has catastrophic impacts on terrestrial ecosystems. It is challenged to quantitatively understand the vegetation vulnerability under compound dry-hot extremes. In this study, a vine-copula conditional probability (VCCP) model is proposed to quantify the impacts of dry-hot events on vegetation dynamics, where the dependence patterns of the Normalized Difference Vegetation Index (NDVI), standardized precipitation evapotranspiration index (SPEI), and standardized temperature index (STI) are modelled through vine copula functions. The VCCP model can evaluate the conditional probability of vegetation loss under multiple dry-hot events and reveal the temporal and spatial patterns of vegetation vulnerability of different land-use types. Then, the VCCP model is applied to Xinjiang province, where the ecological environment is fragile and soil erosion is serious. The dependence patterns among NDVI, SPEI and STI in summer season (June-August) during 1983-2015 are identified. The main findings are: (i) spatial and temporal responses of vegetation to drought and hot events present distinctively; (ii) under the extreme scenario, the average probability of vegetation loss below the 50th percentile in August reach 58.2%, followed by July (with 44.0%) and June (with 33.1%); (iii) the northern and southwestern regions of Xinjiang (especially for the grassland in the mountain areas) have the worst resistance to extreme dry-hot events in summer season. The findings can provide insights into the impacts of compound extremes on vegetation conditions and help decision makers take effective and efficient ecosystem management to mitigate climatic disasters.