The efficacy of different vaccine preparations was investigated for the development of a practical vaccine against pseudotuberculosis in cultured yellowtail. The vaccines were: formalin killed Pasteurella piscicida, a whole cell vaccine; phenol-water extracted lipopolysaccharide (LPS), a LPS vaccine; and an autolysed centrifuged 75% (NH4)2SO4 precipitated culture, a precipitated vaccine. The vaccines were delivered by immersion and spray methods. An artificial challenge by gastral administration with virulent P. piscicida was carried out three weeks after vaccination. The percent survival ten days after challenge in the fish vaccinated by the immersion method were: 87% in group administered the LPS vaccine, 73% using the precipitated vaccine, 40% using the whole cell vaccine and 40% in the non vaccinated controls. The percent survival ten days after challenge in the fish vaccinated by the spray method were: 80% in the group administered the LPS vaccine, 60% in the precipitated vaccine group, 50% in the whole cell vaccine group and 30% in the non-vaccinated controls. Prior to challenge, higher agglutinating antibody titers against P. piscicida in the skin mucus were observed in vaccinated fish that had greater protection after challenge. It was concluded the LPS from P. piscicida cells was the most effective vaccine of the three vaccines examined against pseudotuberculosis in yellowtail.