To address the problems of strong subjectivity in weighting by the analytic hierarchy process and insufficient research on the spatiotemporal distribution of life loss risks in traditional dam breach life loss assessment, as well as the high fuzziness and randomness of risk indicators, a risk assessment method for dam breach life loss based on a combined weighting-cloud model was proposed. This method comprehensively considered various influencing factors, established a grading standard for dam breach life loss assessment indicators, and constructed a unit population distribution model for the downstream flooded area based on dam breach flood routing results. By integrating the indicator weights determined by the combined G1-entropy weight method with the risk membership degree of each unit obtained from the cloud model, dynamic risk assessment of different spatiotemporal units during the dam breach process was achieved. The overall life loss risk level of the reservoir dam breach was determined by calculating the life loss value. The case study results demonstrate that the combined G1-entropy weight method integrates data changes during flood routing and effectively balances the subjective and objective weights of relevant indicators in dam breach life loss risk assessment; the combined weighting-cloud model method can effectively handle the randomness and fuzziness issues in the risk assessment process, achieve dynamic risk assessment of the dam breach process, better conform to the actual flood routing process, and more accurately reflect the impact of the flood on the affected population during the initial stage of the dam breach.