To clarify the microbial community structure characteristics, distribution of potential pathogenic microorganisms, and key influencing factors in drinking water plants in Guangzhou, this study analyzed the microbial community and risk in drinking water plants in Guangzhou by combining high-throughput sequencing technology with culturable methods. Significant differences in microbial communities between Water Plants A and B were revealed using high-throughput sequencing results. The dominant genera in Water Plant A were hgcI_clade (0~30.54%) and Phreatobacter (0~46.36%), whereas the dominant genera in Water Plant B were hgcI_clade (0~46.33%) and Stenotrophomonas (0~46.68%). Potential pathogenic bacteria were further analyzed, and Phreatobacter (46.36%) and Sphingomonas (32.98%) proportions were found to be high in Water Plant A after disinfection. In contrast, Stenotrophomonas (46.70%) and Rhodococcus (41.40%) were the dominant pathogenic bacteria in Water Plant B. Results from culture methods indicated higher microbial diversity and more pathogenic microorganisms in Water Plant B than in Water Plant A. Pathogenic bacteria such as Bacillus cereus and Pseudomonas aeruginosa were isolated from the filter media, suggesting that disinfection, cleaning, or replacement of the filter media should be strengthened in the future. Canonical correlation analysis indicated that temperature, pH, dissolved oxygen, chemical oxygen demand, and total carbon were the main factors affecting the microbial community structure of the drinking water treatment system. In conclusion, both drinking water plants A and B have varying degrees of pathogenic microbial contamination. The key contamination points in the drinking water treatment system should be further controlled in the future. The results of this study provide a theoretical basis for the efficient control of microorganisms in drinking water plants and optimization of treatment processes, and have significant practical significance for ensuring the microbial safety of drinking water.