In petroleum-contaminated aquifers, biodegradation is always associated with various types of microbial metabolism. It can be classified as autotrophic (such as methanogenic and other carbon fixation) and heterotrophic (such as nitrate/sulfate reduction and hydrocarbon consumption) metabolism. For each metabolic type, there are several key genes encoding the reaction enzymes, which can be identified by metagenomics analysis. Based on this principle, in an abnormally low dissolved inorganic carbon (DIC) petroleum-contaminated aquifer in North China, nine groundwater samples were collected along the groundwater flow, and metagenomics analysis was used to discover biodegradation related metabolism by key genes. The major new finding is that autotrophic metabolism was revealed, and, more usefully, we attempt to explain the reasons for abnormally low DIC. The results show that the methanogenesis gene, Mcr, was undetected but more carbon fixation genes than nitrate reduction and sulfate genes were found. This suggests that there may be a considerable number of autotrophic microorganisms that cause the phenomenon of low concentration of dissolved inorganic carbon in contaminated areas. The metagenomics data also revealed that most heterotrophic, sulfate, and nitrate reduction genes in the aquifer were assimilatory sulfate and dissimilatory nitrate reduction genes. Although there was limited dissolved oxygen, aerobic degrading genes AlkB and Cdo were more abundant than anaerobic degrading genes AssA and BssA. The metagenomics information can enrich our microorganic knowledge about petroleum-contaminated aquifers and provide basic data for further bioremediation.