Groundwater-Sensitive Agroecosystem (GWSA) in arid regions are increasingly vulnerable to compound drought-waterlogging disasters driven by groundwater fluctuations under climate change. This study proposes a remote sensing-based framework to monitor such compound stress in the Gahai Irrigation District, Northwest China. By integrating a 1 km-resolution root-zone soil moisture (RZSM) dataset with 33 downscaling factors (e.g., vegetation indices, topography, and hydrological variables), we generated a 30 m-resolution daily RZSM product (2018-2022) using a Random Forest algorithm. The Soil Moisture Condition Index (SMCI-index) and Double Stress Index (DSI) were developed to identify drought (SMCI-index < 0.4 for >= 10 days) and waterlogging (SMCI-index > 0.6 for >= 3 days) events and their synergistic impacts. Results revealed significant spatial heterogeneity in Compound Drought-Waterlogging Stress: moderate stress dominated central GWSA (13.82 km(2)), driven by waterlogging-induced soil degradation, while severe stress (6.04 km(2)) occurred along boundaries with alternating drought-waterlogging dominance. Groundwater level, precipitation, and temperature were key drivers, with temperature paradoxically reducing drought areas via snowmelt-enhanced recharge. Validation of the SMCI-index derived from RZSM data showed strong consistency with vegetation indices (POD > 80 %, r = 0.98) and a 69.27 % spatial overlap with modeled waterlogging zones. This study demonstrates that the use of downscaled RZSM data can effectively mitigate precipitation interference and enable fine-scale monitoring of groundwater-driven compound drought-waterlogging stress. The findings offer critical insights for enhancing agricultural resilience and thus maintaining ecosystem services in GWSA under hydrological extremes.