The urban water ecosystems, such as the landscape ponds are commonly considered under the influence of anthropogenic disturbances, which can lead to the deterioration of the water quality. The prokaryotic communities are considered as one of the best indicators of the water quality. However, there are significant gaps in understanding the ecological processes that shape the composition and function of prokaryotic communities in the urban water ecosystems. Here, we investigated the biogeographic distribution of prokaryotic assemblages in water environments including landscape ponds, drinking water reservoirs, influents (IFs) and effluents (EFs) of wastewater treatment plants of a coastal city (Xiamen), China, by using 16S rDNA amplicon sequencing. Our results indicated that the ponds had higher α-diversity of prokaryotic communities than those in the reservoirs, while there were significant variations in the community compositions among ponds, reservoirs, IFs and EFs. Moreover, ponds harbored a significantly higher proportion of sewage- and fecal-indicator taxa than those in the reservoirs, suggesting the occurrence of exogenous pollution in the urban ponds. Null model analysis revealed that dispersal limitation was the main ecological processes resulting in the divergence of prokaryotic community compositions between ponds and other environments, while dispersal limitation and variable selection played an essential role in the formation of unique prokaryotic assemblages in the reservoirs. Function predication analysis demonstrated that the ponds shared more similar functional profiles with IFs or EFs (e.g., chemoheterotrophy, fermentation, chlorate reducers, nitrate reduction and respiration) than the reservoirs, whereas dominance of photoautotrophy was observed in the reservoirs. Overall, this study provides a profound insight of the ecological mechanisms underlying the responses of prokaryotic communities in the urban landscape ponds to the anthropogenic disturbances.