Flash droughts, characterized by rapid declines in soil moisture, have become increasingly frequent and exert sustained pressure on terrestrial carbon and water cycles. However, a practical framework is still lacking to systematically track vegetation ecosystems' coupled structural and functional responses before, during, and after flash-drought onset. We integrate daily observations of the Normalized Difference Vegetation Index (NDVI), gross primary productivity (GPP), water-use efficiency (WUE), and underlying WUE (uWUE) for China from 1982 to 2019, and use these data to develop a high-temporal-resolution Vegetation Flash Drought Response Framework (HTR-VFDRF). The framework characterises three sequential lag windows: the negative anomaly response time, peak point time, and lag recovery time. Furthermore, we applied the KDE method to analyze the temporal patterns of vegetation structural and functional response sensitivity under flash drought stress, while also employing interpretable machine learning and the geographic detector to analyze the multidimensional drivers of land surface, meteorological, and land-atmosphere coupling factors that influence the negative anomaly peak points in vegetation structure and function. Our results show that vegetation ecosystem functional indicators are more sensitive to flash drought than structural indicators. Preceding and following flash drought events, the vegetation ecosystem function daily-scale negative anomaly response curves exhibit a rapid V-shaped decline and subsequent recovery in humid southeastern China, in contrast to a more gradual S-shaped increase and stabilization in cold-dry and semi-arid regions. Negative anomaly peaks in GPP, WUE, and uWUE deepen from arid to humid zones along a climatic aridity gradient, with mixed forests exhibiting the highest functional stability among vegetation types. Mechanistic analysis reveals that across both monsoon and non-monsoon regions of China, flash drought-induced declines in vegetation carbon uptake are primarily driven by synergistic shifts in soil moisture, air temperature, and surface energy fluxes (SSHF and SLHF), which jointly constrain evapotranspiration and photosynthetic efficiency. Moreover, the interaction between flash-drought intensity and bedrock depth explains approximately 70% of the spatial heterogeneity in long-term mean GPP negative anomaly peak points. These findings emphasize the essential role of high-temporal-resolution monitoring for detecting early functional decline and recovery, and demonstrate that the HTR-VFDRF can serve as a scalable method for evaluating global vegetation resilience and vulnerability to flash drought under continued climate warming.