Abstract: Most existing studies on flash droughts have been conducted from the perspective of meteorological or agricultural droughts. In contrast, research from a hydrological drought perspective, such as research based on streamflow or base flow deficits, remains limited. To address this gap, a flash drought assessment method that accounts for the rapid development characteristics and water deficit conditions of streamflow was established on the basis of run theory and the threshold level method. Then, a digital filtering method was introduced for base flow separation to support the identification of flash droughts in base flow processes, thus complementing existing flash drought identification methods. The Ganjiang River basin was selected as a case study area to evaluate the proposed method. The results indicated that the average durations and onset rates of flash drought events identified from the streamflow and base flow series in the Ganjiang River basin from 1981 to 2020 were 8.40 and 9.92 pentads and 20.15 and 24.40%/pentad, respectively. Validation of the changes in key hydro-meteorological variables suggested that flash droughts characterized on the basis of streamflow performed better in capturing the initial stage of soil moisture flash drought transmission into the watershed streamflow process, whereas the base flow characterization method for flash droughts was more reliable for determining whether a flash drought event had been completely terminated within the basin water cycle process.