Abstract: Bacillus thuringiensis (Bt) is a gram-positive and soil-dwelling bacterium widely used as biological pesticide with broad insecticidal spectra. Most Bt strains can produce Cry proteins which have specific activities against insect species of the orders Lepidoptera, Diptera, Coleoptera, Hymenoptera and Nematodes. Because of the specificity against target insects, Bt has been regarded as an environmentally friendly biopesticide with little or no effect on mammals, birds, amphibians, reptiles and most other beneficial insects. In fact, it has so far not been conclusively studied whether Bt was really safe to natural microbial communities in agricultural ecosystems. In this study, the influence of Bt on soil bacterial population and diversity in cotton fields was investigated by using traditional plating method and PCR-DGGE technique after application of two dosages 0.1 kg·hm^-2 (recommended dosage) and 10 kg·hm^-2 (hgih dosage) of Bt agent. Avermectins and water were respectively used as positive and negative controls. Result indicated that 3 days after treatment, Bt had no significant effect on bacterial abundance in soils. Bacterial abundance reached the maximum 3 days after treatment and began to decline thereafter. Then 12 days after treatment, bacterial abundance in soils treated with water, commercial recommended dosage and high dosage Bt was 4.0×10⁷ CFU·g^-1. Bacterial abundance in soils treated with recommended dosage Bt was not significantly different from that treated with water. Sixth days into the experiment, there was a significantly high bacterial abundance in soils treated with recommended dosage Bt. There was also no significant difference in soil bacterial abundance between high dosage Bt and water during the whole experiment. Six days after treatment, bacterial abundance of soil treated with avermectins was not significantly different from that treated with water. However, bacterial abundance of avermectins treatment was significantly lower than that of the other 3 treatments 12 45 days after treatment. DGGE spectrogram analyses showed that Bt agent had no significant harmful effects on cotton rhizosphere soil bacterial community structures. Based on analyses of bacterial structures of different treatments, electrophoresis bands of samples treated with given dosages Bt and negative control were in the same branch cluster after 12 days of treatment. This indicated that Bt agent had no influence on bacterial community structure after 12 days. The diversity indexes of Bt treatments were same as that of negative control. Compared with Bt agent, avermectins significantly suppressed bacterial abundance in cotton rhizosphere soils. Following DGGE analysis, 17 different sequences (B1 B17) retrieved from the GenBank using BLAST program were isolated. Sequences with 100% homology of Bt were observed only in Bt treatment dosages of 10 kg·hm^-2. This suggested that Bt could not replace native species in agricultural ecosystems if applied at commercial recommended dosage. In addition, beneficial-role bacteria in soils (e.g., Bradyrhizobium sp. which helped Leguminosae to fix nitrogen into soils; Sphingobium sp., Sphingomonas sp. and Rhodococcus sp. which acted as pollution remediation species) were retrieved from the GenBank. Bt had no influence on these functional species in soils at the given dosages. In conclusion, the study provided an evaluation of the effects of Bt on microbial population and community structure using culture-dependent/independent techniques. The study suggested that Bt biocontrol agent was environmentally friendly with no significant effect on natural bacterial communities in soil ecologies.