<p>Climate warming is increasing the frequency and severity of heatwaves and droughts, with important consequences for vegetation productivity and the terrestrial carbon sink. In particular, compound atmospheric–soil drought may constrain ecosystem carbon uptake more strongly than either stressor alone, yet its effects on vegetation productivity have rarely been quantified within a probabilistic framework. Here, we develop a vine copula-based conditional probability approach to assess the vulnerability of vegetation productivity to compound drought across China from 2002 to 2022, using soil moisture (SM) and vapor pressure deficit (VPD) as indicators of soil and atmospheric drought, respectively. We find the following: (1) gross primary productivity (GPP) increased significantly across 60.14% of China, whereas VPD rose mainly in the southwest and declined in parts of the northeast and northwest; SM showed mixed regional trends, with 18.19% of areas increasing and 23.37% decreasing. (2) GPP was predominantly positively associated with SM and negatively associated with VPD, while SM and VPD exhibited a strong and spatially consistent negative correlation, indicating marked synergy between soil and atmospheric drought. (3) Under compound drought, the probability structure of productivity risk showed a clear pattern of higher vulnerability in northern China and lower vulnerability in the south, with hotspots concentrated in the arid and semi-arid transition belt. As drought intensity increased, the probability of low productivity rose sharply and expanded spatially, revealing strong tail amplification, with grasslands showing the greatest sensitivity.</p>