Abstract: In order to deeply understand the mechanism of basin floods under the extreme rainfall condition, taking the "23∙7" catastrophic flood occurred at the Daqinghe River as an example, a distributed modelling-based quantitative analysis approach to assess the spatiotemporal characteristics of basin flood is proposed in this study. The spatiotemporal variable-source hydrological simulation method has been taken to set up a distributed model consisted with 19 computation units covering the south and north branches of the Daqinghe River, and to run in parallel. The average flood peak error of the simulation results at 14 calibrated stations is lower than 5%, and the average Nash coefficient is 0.7. The results show that the averaged runoff coefficient of the north branch of the Daqinghe River is 0.55, which is much higher than that of the south branch by 0.27. The flow amount which causes this flood disaster mainly comes from the north branch. During the "23∙7" catastrophic flood process, the north basin exhibits a mixed pattern of Dunne and Horton runoff generation mechanism. When the accumulated rainfall in the basin reaches 107mm, the Horton surface runoff will be appeared in the catchment. When the rainfall intensity exceeds 15.3mm/h, the Dunne surface runoff will be occurred. Meanwhile, compared to the south branch of the Daqinghe River, the lack of large-scale controlling projects in the main branch of north branch is another major reason led to this catastrophic flood disaster. The approach presented in this paper can provide technical references for the construction of flood model integrated in the digital twin system of other basins in China.