<p>Evaluating the spatiotemporal characteristics of flood hazard risk is essential for stakeholders to undertake disaster preparedness and mitigation activities. Traditional hydrological modeling, hydrological station methods, and multi-indicator analysis techniques struggle to objectively and accurately capture the spatiotemporal characteristics of flood disaster risk due to challenges in spatial calibration and sparse spatial distribution, respectively. Historical remote sensing data can identify areas actually affected by floods, thereby improving flood risk assessment. In the study, Landsat and Sentinel remote sensing data from 1990 to 2023 were utilized to delineate flood inundations in the Poyang Lake region. Employing the Flood Exceedance Probability as an indicator, the study analyzes the spatiotemporal dynamics of flood disaster risk. Results indicate that: (1) Flood disaster risk is higher on the west, southwest, south, and southeast sides of Poyang Lake, along rivers, and in urban built-up areas. (2) Since 2000, the flood disaster risk in the Poyang Lake region has significantly decreased. (3) About 9.36% of the region has seen a reduction in flood risk, mainly around Poyang Lake and along rivers, related to improved disaster prevention and mitigation capabilities; about 3.80% has experienced increased risk, primarily in densely populated and built-up areas, associated with urban expansion and population growth. This study, for the first time, employs long-term earth observation data to objectively analyze the spatiotemporal variations of flood disaster risk in the Poyang Lake region, unveiling its temporal and spatial trends. These findings hold significant practical value for stakeholders in advancing flood prevention and mitigation efforts.</p>