Brucella melitensis and Brucella ovis are gram-negative pathogens of sheep that cause severe economic losses and, although B. ovis is non-zoonotic, B. melitensis is the main cause of human brucellosis. B. melitensis carries a smooth (S) lipopolysaccharide (LPS) with an N-formyl-perosamine O-polysaccharide (O-PS) that is absent in the rough LPS of B. ovis. Their control and eradication require vaccination, but B. melitensis Rev 1, the only vaccine available, triggers anti-O-PS antibodies that interfere in the S-brucellae serodiagnosis. Since eradication and serological surveillance of the zoonotic species are priorities, Rev 1 is banned once B. melitensis is eradicated or where it never existed, hampering B. ovis control and eradication. To develop a B. ovis specific vaccine, we investigated three Brucella live vaccine candidates lacking N-formyl-perosamine O-PS: Bov::CA¿wadB (CO2-independent B. ovis with truncated LPS core oligosaccharide); Rev1::wbdR¿wbkC (carrying N-acetylated O-PS); and H38¿wbkF (B. melitensis rough mutant with intact LPS core). After confirming their attenuation and protection against B. ovis in mice, were tested in rams for efficacy. H38¿wbkF yielded similar protection to Rev 1 against B. ovis but Bov::CA¿wadB and Rev1::wbdR¿wbkC conferred no or poor protection, respectively. All H38¿wbkF vaccinated rams developed a protracted antibody response in ELISA and immunoprecipitation B. ovis diagnostic tests. In contrast, all remained negative in Rose Bengal and complement fixation tests used routinely for B. melitensis diagnosis, though some became positive in S-LPS ELISA owing to LPS core epitope reactivity. Thus, H38¿wbkF is an interesting candidate for the immunoprophylaxis of B. ovis in B. melitensis-free areas.