<p>Flash floods are characterized by their destructive power, rapid onset, and unpredictability, often causing severe damage to both natural environments and socioeconomic systems. Understanding the detailed disaster-causing mechanisms of flash floods is critical for effective disaster risk reduction. However, current studies have not captured the comprehensive circumstance of flash floods that integrates environment, hazard, and exposure from the perspective of disaster systems theory. To address the gap, this study established a systematic framework for comprehensively evaluating flash floods disaster-causing mechanisms in ungauged mountainous micro-watersheds by integrating multi-source data, including remote sensing observations, meteorological station data, unmanned aerial vehicle measurements, and participatory geographic information system data, with hydrological-hydrodynamic and statistical models. The proposed framework consists of four interconnected steps: design storm estimation, flash flood process simulation, critical rainfall calculation, and disaster loss evaluation. Through a case study conducted in Qialegeer Village, Xinjiang, China, we demonstrated the framework’s applicability by reconstructing flash flood scenarios, including the 2017 event as well as those of 10 and 20&#xa0;years return periods. The results demonstrate that our framework robustly and systematically elucidates flash flood disaster process in the region with high reliability. Furthermore, it is adaptable to other ungauged mountainous micro-watersheds. This framework ultimately serves to enhance disaster risk mitigation and build resilience in vulnerable mountainous communities.</p>