Flash droughts are characterized by an abrupt onset and swift intensification. Global surface soil moisture (θRS) from NASA’s Soil Moisture Active Passive (SMAP) satellite can facilitate a near-real-time assessment of emerging flash droughts at 36-km footprint. However, a robust flash drought monitoring using θRS must account for the i) short observation record of SMAP, ii) non-linear geophysical controls over θRS dynamics, and, iii) emergent meteorological drivers of flash droughts. We propose a new method for near-real-time characterization of droughts using Soil Moisture Stress (SMS, drought stress) and Relative Rate of Drydown (RRD, drought stress intensification rate) ─ developed using SMAP θRS (March 2015-2019) and footprint-scale seasonal soil water retention parameters and land-atmospheric coupling strength. SMS and RRD are nonlinearly combined to develop Flash Drought Stress Index (FDSI) to characterize emerging flash droughts (FDSI ≥ 0.71 for moderate to high RRD and SMS). Globally, FDSI shows high correlation with concurrent meteorological anomalies. A retrospective evaluation of select droughts is demonstrated using FDSI, including a mechanistic evaluation of the 2017 flash drought in the Northern Great Plains. About 5.2% of earth’s landmass experienced flash droughts of varying intensity and duration during 2015-2019 (FDSI ≥ 0.71 for >30 consecutive days), majorly in global drylands. FDSI shows high skill in forecasting vegetation health with a lead of 0-2 weeks, with exceptions in irrigated croplands and mixed forests. With readily available parameters, low data latency, and no dependence on model simulations, we provide a robust tool for global near-real-time flash drought monitoring using SMAP.