[Objective] For this paper a pot experiment was conducted to evaluate effects of water management varying in pattern on composition of soil selenium, bacterial community diversity in rhizosphere soil and Se accumulation in grains of rice, and further to explore mechanism of water management affecting the absorption and accumulation of selenium in rice, so as to provide considerable theoretical and practical support to effective regulation of selenium transfer from rice field to agricultural products, to meet the demand of human for selenium nutrient and health.[Method] The experiment was designed to have three treatments in water management treatments, that is, (I) flooded irrigation(F); (II) aerobic irrigation(A); and (III) alternation of flooded and aerobic irrigation (AFA), and three replicates for each treatment. Starting from D20 (20 days after rice transplanting), soil pH and Eh were measured every 12 days. The selenium in the rhizosphere soil was fractionated at the tillering stage, heading stage and maturity stage of the crop. Content of selenium in rice roots, stems, leaves and grains as well as yield of the crop were measured at maturity. Before rice harvest, the fresh soil in the rhizosphere bag was retrieved and placed in a bag, which was then sealed and stored at low temperature (-80℃) for high-throughput sequencing analysis.[Result] Results show that at all the rice growth stages, Treatments A and AFA were higher than Treatment F in soil pH and Eh to a certain extent and in content of soluble and exchangeable selenium as well, thereby in availability of soil selenium. In terms of selenium content, the organs of a rice plant exhibited an order of root (1.38-2.22 mg·kg^-1) > leaf (0.55-0.85 mg·kg^-1) > stem (0.53-0.61 mg·kg^-1) > grain (0.15-0.53 mg·kg^-1). Treatment AFA was the highest in selenium content in rice grains, while Treatment F was the lowest. The difference between the two was obvious. Treatment AFA was significantly or 7.83% higher than Treatment F and 13.51% higher than Treatment A in grain yield. The dominant species of bacteria in the rhizosphere soil of rice were Proteobacteria, Chloroflexi, Bacteroidetes, Acidobacteria, Patescibacteria and Gemmatimonadetes, with Proteobacteria being the most abundant one in rice rhizosphere soil in all the treatments. Obviously water management had a significant impact on richness of the bacteria. Treatments AFA and A were significantly higher than Treatment F in richness of Proteobacteria. On the level of class, abundance of Gammaproteobacteria was positively related to content of available Se in the soil, so the increase in the former may be deemed as an important reason for higher bioavailability of selenium in the soil.[Conclusion] Comprehensive analysis shows that alternation of flooded and aerobic irrigation can not only improve yield and selenium content of rice grains, but also save water for irrigation as compared to normal flooding irrigation. Therefore it can be recommended as preferential water management method in rice production.