Abstract: Denitrification is critical for nitrogen cycle in the ecosystem, where fixed nitrogen is released into the atmosphere as N₂. Nitrite reductase, the product of nirS or nirK nitrite reductase genes, is the key enzyme of bacteria dissimilatory denitrification process. Denitrifying bacteria community composition varies with environmental factors such as temperature, moisture, pH, O₂ and nutrient availability. There is obvious denitrification process in flooded paddy fields. Hence denitrifying bacteria community structure and abundance in paddy fields are used to investigate the response of denitrifying bacteria to nitrogen fertilizer application in paddy fields. The experiment was conducted in a second-year nitrogen fertilization field with the aid of denaturing gradient gel electrophoresis and real-time PCR assay copies of nirS gene. DGGE images of nirS gene in root-zone soil and surface soil showed rich abundance of denitrifying bacteria in paddy soils. DGGE band number in surface soil image was higher than that in root-zone soil. Principle components analysis (PCA) of nirS gene DGGE profile showed that denitrifying bacteria community structure in root-zone or surface soil of paddy fields with nitrogen fertilizer N: 150 kg(N)·hm^-2 was similar to that of paddy fields without fertilizer (CK) during rice growth stages of tillering, heading and maturity. Also no difference was noted in denitrifying bacteria community structure in root-zone soil or surface soil among different growth stages of rice. Denitrifying bacteria nirS gene copy abundance in root-zone or in surface soil with nitrogen fertilizer treatment was significantly (P < 0.05) higher than that of CK treatment during rice growth. In both nitrogen fertilizer and CK treatments, denitrifying bacteria nirS gene copies in root-zone soil markedly (P < 0.05) dropped at maturity stage of rice growth. There were, however, no differences in nirS gene copies in surface soil among the different rice growth stages. At maturity stage, nirS gene copies in surface soils of both nitrogen fertilizer and CK treatments were higher (P < 0.05) than those in root-zone soils. Furthermore, rice yield in nitrogen fertilizer treatment was 44% higher than that of CK. In conclusion, denitrifying bacteria abundance was not only variable but also actively responded to nitrogen fertilizer supply. On the other hand, denitrifying bacteria community structure was not only relatively stable but largely unresponsive to nitrogen fertilizer supply. The study demonstrated that nitrogen fertilizer enhanced denitrifying bacteria abundance, which was critical for nitrogen cycling in paddy field ecosystem.