In this study, a novel numerical model is first developed, which can simulate compound floods under a framework comprehensively considering the combined effects of tide, river flow, rainfall, and wind. The modelling framework is applied to reproduce an extreme compound flood event in the Pearl River Delta caused by Typhoon Hato (2017). The model is estimated to perform reasonably well, which can yield consistent results compared with observations including winds, water levels, and inundation depths over the study area. Inundation was concentrated on downstream estuaries and floodplains and coastal forcings played a dominant role in this compound event. However, ignoring river and rainfall contributions would result in an underestimation of flood extent and water levels that could be up to 48%. Sensitivity tests indicate that precise calibration of the roughness coefficient is more critical in river and estuarine environments than in land region. This study demonstrates the importance of utilizing an integrative framework to assess and predict compound flooding hazards driven by the effects of tropical cyclones.