Soybeans grown in Africa have been selected to nodulate effectively with indigenous Bradyrhizobium spp. populations since Bradyrhizobium japonicum populations are considered absent or in very low numbers in African soils. The major objective of this study was to estimate total population of Bradyrhizobia specific to soybean in two agro-ecologically different study sites, Kiboko in Makueni District, Southeast Kenya (semi-arid to arid conditions) and Kaguru in Meru District, East Kenya (semi-humid climate) . The population of the indigenous rhizobia specific to soybeans was determined using the Most Probable Number (MPN) plant infection technique. In these experiments, the total Bradyrhizobia populations, the population sizes of taxonomically defined slow-growing Bradyrhizobia specific to soybean and the population sizes of Bradyrhizobia spp. specific to tropical Glycine Cross (TGx) varieties were determined for the two study sites. Cowpea, Vigna unguiculata, cultivar Ken Kunde I was used to estimate the total Bradyrhizobia spp. population. Clark soybean, Glycine max, was used to estimate the population sizes of taxonomically defined slow-growing Bradyrhizobia spp. specific to soybean while a TGx genotype, SB12-TGx1869-31E was used to determine the population sizes of Bradyrhizobia spp. specific to TGx varieties. The results of the MPN counts indicated that the total Bradyrhizobia population in Kiboko was between 2.59x104 and 1.89x105. The population size of taxonomically defined slowgrowing Bradyrhizobia in Kiboko was between 2.59x102 and 1.89x103 cells per gram of soil sample while the approximate Bradyrhizobia population specific to TGx genotype was between 7.81x102 and 5.67x103 cells per gram of soil. In Kaguru, the approximate total Bradyrhizobia population was between 1.04x102 and 7.56x103 cells per gram of soil. The population size of taxonomically defined slow-growing Bradyrhizobia was be-tween 1.33x102 and 9.72x102 cells per gram of soil while the approximate Bradyrhizobia population specific toTGx genotype was between 2.37x102 and 1.73x103 per gram of soil. These populations were adequate to give satisfactory results on nodulation and nitrogen fixation in the two study sites.