As climate change intensifies, extreme weather events, especially the coastal flooding is becoming increasingly severe. Carbon emissions exacerbate climate change, leading to increased frequency and intensity of extreme precipitation, which may result in flooding. This study focuses on the Guangdong-Hong Kong-Macao Greater Bay Area (GBA), a coastal city cluster in China, and adopts the scenario analysis approach to evaluate flood risks in the different carbon mitigation scenario from the perspectives of dangerousness, and adaption ability. The result shows that the Huizhou in the east of the GBA and the Zhaoqing in the northwest have the highest level of flood disaster risk. However, the central region of the GBA (Guangzhou, Dongguan, Zhongshan, and Foshan) is very vulnerable to disaster-inducing environments and hazard-affected bodies, which means that the flooding risk and the impact of such risk in the GBA display regional differentiation characteristics and the spatial patterns of them are different. According to the carbon neutrality scenario, the flood risk in the GBA peaks in 2030, and decreases during 2030–2090. In contrast, under the uneven development and high carbon pathways, the risk peaks in 2060, and the overall risk by 2090 is roughly the same as that in 2060. In addition, the regions with the highest flood disaster risk are Guangzhou, Foshan, Dongguan and Zhongshan along the Pearl River estuary. Overall, it shows that areas with higher levels of urbanization and more developed economies will face higher flooding risks. The “dangerousness-adaption ability” structure can reflect the spatial distribution pattern of flood disaster risks, identify the priority areas for flood prevention and control, and is conducive to risk management. At the same time, the research results show that ecological protection have a positive effect on reducing flood risk. In addition, corresponding policy recommendations for pre-disaster, mid-disaster and post-disaster have been proposed, providing references for flood risk management in the GBA.