Bio-prospecting was conducted in Kenya to identify elite isolates of rhizobia capable of effectively nodulating promising soybean varieties. One hundred isolates were recovered from nodules of wild and cultivated legume hosts. These isolates were authenticated and tested for effectiveness on soybean (Glycine max) var. SB 19 in sterile vermiculite, and the twenty-four most promising isolates screened in potted soil to assess their competitive abilities on two varieties ("promiscuously nodulating" SB 19 and specific SC Safari). The six best performing isolates were then evaluated under field conditions, comparing them to strain USDA110. Test isolates were classified into; non-infective, ineffective, partly effective, effective and highly effective based on their performance relative to controls and industry standards. In potted soil, native rhizobia isolates nodulated promiscuous soybean (SB19) but only 46% of them nodulated specific soybean (Safari). In the field experiment, isolate NAK 128 performed best on both promiscuous and specific soybean varieties, significantly (p<0.05) outperforming USDA110 by 29% and 24%, respectively. Partial economic return to inoculation with NAK 128 was about 21:1, justifying inoculation as a field practice and producing up to 2.5 million nodules (334 kg) ha-1, significantly (p<0.05) more than USDA 110. The best isolates from this investigation have commercial potential.