Brucellosis is a serious zoonosis that occurs worldwide, and its diagnosis is typically based on the detection of antibodies against Brucella lipopolysaccharide (LPS). However, the specificity of the LPS-based test is compromised by cross-reactivity with Escherichia coli O157:H7 and Yersinia enterocolitica O:9. Also, diagnosis based on the LPS test cannot differentiate between vaccinated and infected individuals. The detection of the 26-kDa cytosoluble protein (BP26) antibody is considered an alternative that circumvents these drawbacks because it is exclusively expressed by infectious Brucella. A BP26-based enzyme-linked immunosorbent assay (ELISA) has been tried for the diagnosis of Brucella-infected animals and humans, but a few results showed that BP26 couldn't react with all Brucella-positive sera. In order to explore whether different animals could produce antibodies against BP26 after being infected with various Brucella species, we infected sheep, goats, and beef cattle with common virulent reference Brucella species. All sera were collected from the experimental animals and tested using both LPS-based ELISAs and BP26-based ELISAs. The results showed that all Brucella-infected individuals could produce high levels of antibodies against LPS, but only B. melitensis 16M- and B. melitensis M28-infected sheep and B. melitensis 16M- and B. abortus 2308-infected goats could produce antibodies against BP26. Therefore, we concluded that the BP26-based indirect ELISA (i-ELISA) showed both Brucella species and host specificity, which obviously limits its reliability as a substitute for the traditional LPS-based ELISA for the detection of brucellosis.