The occurrence process of flood disasters has a certain degree of predictability. Predictive analysis of flood risk can mitigate or reduce the impact of disasters and improve disaster prevention and reduction capabilities. The analysis of flood hazard warning is an important foundation for disaster prevention and reduction. Early warning before the occurrence of disasters can effectively reduce the impact of disasters. The research is focused on flood disasters in the Yangtze River-Huaihe River basin during June to August 2020. This paper aims to develop an improved flood hazard risk warning analysis model using the Analytic Hierarchy Process (AHP) and multiple triggering and predisposing factors. Triggering factors, including cumulative precipitation in the previous three days, current soil moisture, and forecasted precipitation, are crucial in assessing the immediate risks of flood disasters. Predisposing factors, such as river network density, terrain elevation, terrain amplitude, and land use data, provide insights into the vulnerability of the region to flood disasters. By combining these factors, we can effectively evaluate the flood risk and issue timely warnings to mitigate the impact of disasters. To validate the effectiveness of the proposed model, we compare the evaluation results with the flood disaster information reported in the “Meteorological Disaster Management System” of the China Meteorological Administration. The evaluation accuracy rate, which measures the agreement between the risk assessment and the actual occurrence of disasters, reaches 74.46%. This indicates that the model has a relatively high accuracy in predicting flood risks. Additionally, the missing rate, which measures the proportion of missed warnings, is only 5.59%, demonstrating the model’s ability to effectively capture potential flood disasters. Furthermore, the evaluation results show a good correlation between the risk assessment and the actual occurrence of disasters. The warning rate of the maximum disaster unit index, which represents the highest risk within a county, reaches 81.6%. Moreover, for “extreme” heavy rain and flood disasters, the warning rate exceeds 77.3%. This suggests that the proposed model is particularly effective in predicting and warning against severe flood disasters. In terms of temporal consistency evaluation, the risk index consistently increases 3-5 days before the occurrence of “extreme” heavy rain and flood disasters. In conclusion, the model’s high accuracy and reliability make it a valuable tool for decision-making in disaster prevention and reduction efforts. By providing timely and accurate warnings, the model can significantly mitigate the impact of flood disasters and improve the region’s resilience to such events. Future research can focus on further refining the model and incorporating additional factors to enhance its predictive capabilities.