Soil pollution caused by polycyclic aromatic hydrocarbons (PAHs) is threatening human health and environmental safety. Investigating the relative prevalence of different PAH-degrading genes in PAH-polluted soils and searching for potential bioindicators reflecting the impact of PAH pollution on microbial communities are useful for microbial monitoring, risk evaluation, and potential bioremediation of soils polluted by PAHs. In this study, three functional genes, pdo1, nah, and C12O, which might be involved in the degradation of PAHs from a coke factory, were investigated by realtime quantitative PCR (qPCR) and clone library approaches. The results showed that the pdo1 and C12O genes were more abundant than the nah gene in the soils. There was a significantly positive relationship between the nah or pdo1 gene abundances and PAH content, while there was no correlation between C12O gene abundance and PAH content. Analyses of clone libraries showed that all the pdo1 sequences were grouped into Mycobacterium, while all the nah sequences were classified into three groups: Pseudomonas, Comamonas, and Polaromonas. These results indicated that the abundances of nah and pdo1 genes were positively influenced by levels of PAHs in soil and could be potential microbial indicators reflecting the impact of soil PAH pollution and that Mycobacteria were one of the most prevalent PAHs degraders in these PAH-polluted soils. Principal component analysis (PCA) and correlation analyses between microbial parameters and environmental factors revealed that total carbon (TC), total nitrogen (TN), and dissolved organic carbon (DOC) had positive effects on the abundances of all PAH-degrading genes. It suggests that increasing TC, TN, and DOC inputs could be a useful way to remediate PAH-polluted soils.