Waterborne pathogens have attracted a great deal of attention in the public health sector over the last several decades. However, little is known about the pathogenic microorganisms in urban water systems. In this study, the bacterial community structure of 16 typical surface waters in the city of Beijing were analyzed using Illumina MiSeq high-throughput sequencing based on 16S rRNA gene. The results showed that Bacteroidetes, Proteobacteria and Actinobacteria were the dominant groups in 16 surface water samples, and Betaproteobacteria, Alphaproteobacteria, Flavobacteriia, Sphingobacteriia and Actinobacteria were the most dominant classes. The dominant genus across all samples was Flavobacterium. In addition, fifteen genus level groups of potentialy pathogenic bacteria were detected within the 16 water samples, with Pseudomonas and Aeromonas the most frequently identified. Spearman correlation analysis demonstrated that richness estimators (OTUs and Chao1) were correlated with water temperature, nitrate and total nitrogen (p < 0.05), while ammonia-nitrogen and total nitrogen were significantly correlated with the percent of total potential pathogens (p ≤ 0.05). These results could provide insight into the ecological function and health risks of surface water bacterial communities during the process of urbanization.