The activated sludge of a biochemical unit (WLK_OD) and an advanced denitrification unit (WLK_AD) were collected from a municipal wastewater treatment plant (WWTP), in which the TN concentration of effluent was less than 1.5 mg·L -1 , and their microbial community structure and function profiles were analyzed using 16S rRNA gene high-throughput sequencing. The microorganisms in WLK_AD had lower evenness compared with that in WLK_OD, which was attributed to environmental selection. Furthermore, PCoA revealed that different incoming wastewaters had an impact on microbial community structure. At the phylum level, Proteobacteria (70.11%) was enriched in WLK_AD. At the genus level, Thauera, Flavobacterium, Hydrogenophaga , and Zoogloea served as distinct-dominant denitrifying bacteria in WLK_AD; however, Trichococcus (3.50%) and Terrimonas (1.10%) were enriched in WLK_OD. Through the comparison between groups ( P <0.05), the biomarkers detected in each WWTP were different. Furthermore, the results of the co-occurrence network showed that the bacteria from module I had a higher proportion in WLK_AD; the bacteria from module II had a higher proportion in WLK_OD, and they were common microorganisms in WWTPs, implying that wastewater environments drpve the differences in the microbial community structure. Among the types of environmental parameters, the removal efficiency of COD and TN had the greatest impact on the microbial community by the RDA. The removal efficiency of COD was positively correlated with the dominant bacteria from WLK_OD, such as Saccharibacteria, Thermomarinilinea, Terrimonas , and Comamonas ; the removal efficiency of TN was positively correlated with the denitrifying bacteria from WLK_AD, such as Dokdonella, Thauera, Flavobacterium , and Zoogloea . WLK_AD was enriched with Novosphingobium, Dokdonella, Thauera , and Sphingomonas, which synergistically removed TN, leading to the TN of the effluent being less than 1.5 mg·L -1 . Moreover, based on the results of function prediction, WLK_AD had a higher proportion of genes that could code the denitrification enzymes.