Identifying the complex patterns of human-flood interactions over longer periods of time is very important in floodplain management activities. The recently introduced socio-hydrology (SH) model contributes to capture these long-term behaviors of human-flood systems. This model can be utilized to explain the long-term dynamics of human-water interaction in floodplains. The current SH model exclusively illustrates the impact of river floods on floodplain communities. However, in some river basins, urban floods (due to high intensity rainfall) are dominant, whereas in other river basins, both river floods and urban floods influence the dynamics of the system. It is often difficult to distinguish the type of flood from actual local disaster data sets. In this study, we proposed an improvement to the existing SH model to capture the dynamics of both river floods and urban floods based on a case study from the Lower Kelani Basin, Sri Lanka, using simulated historical flood damages. The improved model was applied to capture flood damages in the target watershed, and the results further emphasize the importance of flood risk perception in flood damage reduction.