This study was performed to test the hypothesis that Brucella abortus strain RB51 (SRB51) might protect Korean indigenous mongrel dog against challenge with either virulent B. abortus biotype 1 or B. canis. A total of 12 Korean mongrel dogs were divided into four groups (Group A, B, C and D). Dogs belonging to Group A and C were inoculated subcutaneously with $1{ imes}10^9$ CFU of SRB51 in 1ml of sterile phosphate buffered saline (PBS). Dogs of Group B and D were inoculated subcutaneously with 1ml of sterile PBS as control. At 12 weeks post vaccination, dogs of Group A and B were challenged by oral inoculation of virulent strain of B. canis ($5.0{ imes}10^9$ CFU) and dogs of Group C and D were challenged by oral inoculation of virulent strain of B. abortus biotype 1 ($4.4{ imes}10^{10}$ CFU). The serum antibodies titers in all dogs were monitored at regular interval for eight weeks after challenge (AC) by standard tube agglutination test, plate agglutination test, rose bengal test, 2-mercaptoethanol rapid slide agglutination test and 2-mercaptoethanol tube agglutination test. No antibody titers in Group A and C was detected. Also, the challenge strains were not found from blood of all dogs of Group A and C from 1 week AC till the end of the experiment by culture and modified AMOS-PCR, whereas B. canis and B. abortus challenge strains were detected from blood of Group B and D, respectively. In addition, neither of two challenge bacteria was recovered from liver, spleen, kidneys, lymph nodes and reproductive tracts of Group A and C dogs after postmortem. However, B. canis and B. abortus challenge strains were isolated from these tissues of Group B and D, respectively. These data suggest that SRB51 could be a promising vaccine candidate for immunizing dogs to control canine brucellosis caused by B. canis or B. abortus.