Wastewater treatment plants (WWTPs) rely mainly on the microbial assemblages to contribute significantly for the removal of organic pollutants and nutrients. However, limited information is available on the ecological driving forces underlying the turnover of prokaryotic communities across wastewater treatment processes (i.e., from influents (IFs) and effluents (EFs)) within WWTPs. Here, we used a combination of the 16S rRNA gene amplicon sequencing and a quantitative ecological null model analysis to explore the ecological processes governing the turnover of the prokaryotic communities and the dominant taxonomic taxa across wastewater treatment processes of five full-scale WWTPs in China. Our results indicated that a significant variation in the composition of prokaryotic communities and the dominant taxa between IFs and EFs. The analysis of the environmental sources of indicator OTUs showed that a relatively lower abundance of the sludge/sewage and human guts associated OTUs in EFs than in IFs. Ecological null models revealed that among the ecological processes, deterministic processes were dominant in controlling the turnover of the overall communities from IFs to EFs, whereas the relative importance of deterministic processes varied among the dominant taxa (i.e., Bacteroidetes > Proteobacteria > Gammaproteobacteria > Firmicutes > Betaproteobacteria). However, the assembly of IF and EF communities was influenced mainly by the deterministic and stochastic processes, respectively. In addition, our results indicated that EF communities have a higher phylogenetic diversity than those of the IF communities, but the abundance of prokaryotic 16S rRNA genes was lower in EFs than in IFs. Overall, our study provides a novel insight of the assembly mechanisms underlying the turnover of prokaryotic communities during wastewater treatment processes.