The population diversity and seasonal changes of bacterial communities in rice soils were monitored using both culture-dependent approaches and molecular methods. The rice field plot consisted of twelve subplots planted with two genetically-modified (GM) rice and two non-GM rice plants in three replicates. The DGGE analysis revealed that the bacterial community structures of the twelve subplot soils were quite similar to each other in a given month, indicating that there were no significant differences in the structure of the soil microbial populations between GM rice and non-GM rice during the experiment. However, the DGGE profiles of June soil after a sudden flooding were quite different from those of the other months. The June profiles exhibited a few intense DNA bands, compared with the others, indicating that flooding of rice field stimulated selective growth of some indigenous microorganisms. Phylogenetic analysis of l6S rDNA sequences from cultivated isolates showed that, while the isolates obtained from April soil before flooding were relatively evenly distributed among diverse genera such as Arthrobacter, Streptomyces, Terrabacter, and Bacillus/Paenibacillus, those from June soil after flooding mostly belonged to the Arthrobacter species. Phylogenetic analysis of 16S rDNA sequences obtained from the soil by cloning showed that April, August, and October had more diverse microorganisms than June. The results of this study indicated that flooding of rice fields gave a significant impact on the indigenous microbial community structure; however, the initial structure was gradually recovered over time after a sudden flooding.