Crude oil samples with high- and low-water content from two offshore platforms (PA and PB) in Campos Basin, Brazil, were assessed for bacterial communities by 16S rRNA gene-based clone libraries. RDP Classifier was used to analyze a total of 156 clones within four libraries obtained from two platforms. The clone sequences were mainly affiliated with Gammaproteobacteria (78.2% of the total clones); however, clones associated with Betaproteobacteria (10.9%), Alphaproteobacteria (9%), and Firmicutes (1.9%) were also identified. Pseudomonadaceae was the most common family affiliated with these clone sequences. The sequences were further analyzed by MOTHUR, yielding 81 operational taxonomic units (OTUs) grouped at 97% stringency. Richness estimators also calculated by MOTHUR indicated that oil samples with high-water content were the most diverse. Comparison of bacterial communities present in these four samples using LIBSHUFF and Principal Component Analysis (PCA) indicated that the water content significantly influenced the community structure only of crude oil obtained from PA. Differences between PA and PB libraries were observed, suggesting the importance of the oil field as a driver of community composition in this habitat.