Here, the performance and mechanism of pollutant removal in the Zhoucun reservoir by water-lifting and aeration systems (WLAs) were explored. The hypolimnion anoxic layer disappeared and the reservoir was mixed after the WLAs were operated for approximately 35 days, providing a suitable environment for pollutant removal. Operation of the system enhanced the metabolic activity of the water microbes and their capacity for purification, which contributed to the removal of nitrogen, organic carbon, Fe, Mn, P, and S. Specifically, the total N concentration decreased from 2.55 to 0.48 mg/L, showing an 81.18% removal rate. Microbial metabolism and the diversity index increased following the operation of the WLAs in the Zhoucun Reservoir. Furthermore, the water reservoir clearly inhibited the performance of Fe, Mn, P, and S through the WLA operation, meeting the requirements for class III based on the Chinese Surface Water Environmental Quality Standard (GB3838-2002). High-throughput sequencing analysis revealed increased levels of indicator and keystone operational taxonomic units belonging to Flavobacterium, hgcI_clade, Rheinheimera, Dechloromonas, Pseudomonas, and Rhodocyclaceae, which are related to the degradation of organic carbon and removal of nitrogen and phosphorus. Moreover, total N, ammonia, total P, dissolved oxygen, temperature, and pH were the principal factors affecting the microbial community based on redundancy analysis and the Mantel test. Furthermore, network analysis showed that symbiotic relationships accounted for the major proportion of the microbial network. Our results provide a theoretical foundation for the efficiency of N removal and essential technical support for improving the self-repair capacity of water in drinking water reservoirs.