Flood resilience has aroused significant interest in coastal areas dealing with a growing frequency of severe rainstorms caused by climate change and urbanisation. At the core of flood resilience is the development of a resilient green-grey-blue infrastructure system that can resist, absorb, and recover from floods in a timely manner. Current flood resilience research, however, is limited to evaluating single infrastructure systems, failing to examine the dynamic process or find ideal spatial infrastructure designs for decision-makers. This research proposes a scenario-based assessment framework for integrated green-grey-blue infrastructure systems to improve flood resilience during urban design decision-making. Rainfall-runoff, drainage networks, and river system models are interlinked to provide quantitative simulation evaluations of water quantity and urban impact in various spatial organisations of infrastructure design. A dynamic, multi-criteria decision-making process is used to reveal the importance of five temporal indicators and rank design alternatives. In Guangzhou, China, the efficiency of this architecture is demonstrated on Pazhou Island, a typical river network area. Given the limited water and green space available, the results demonstrate that submerged areas exert a greater influence during peak rainfall, and blue infrastructure storage becomes an essential factor following rainfall. Furthermore, from a spatial perspective, the looped network of green-blue infrastructure enhances flood resilience, and downstream waterway connections and green space-aligned waterways boost the water storage capacity of green-grey-blue infrastructure. This paradigm can improve flood resilience in the Greater Bay Area in the future, especially in response to heavy rainstorms and river floods.