OBJECTIVE To investigate the impact of age and inferred prior vaccination history on the persistence of vaccine-induced antibody against rabies in horses. DESIGN Serologic response evaluation. ANIMALS 48 horses with an undocumented vaccination history. PROCEDURES Horses were vaccinated against rabies once. Blood samples were collected prior to vaccination, 3 to 7 weeks after vaccination, and at 6-month intervals for 2 to 3 years. Serum rabies virus-neutralizing antibody (RVNA) values were measured. An RVNA value of ≥ 0.5 U/mL was used to define a predicted protective immune response on the basis of World Health Organization recommendations for humans. Values were compared between horses < 20 and ≥ 20 years of age and between horses inferred to have been previously vaccinated and those inferred to be immunologically naïve. RESULTS A protective RVNA value (≥ 0.5 U/mL) was maintained for 2 to 3 years in horses inferred to have been previously vaccinated on the basis of prevaccination RVNA values. No significant difference was evident in response to rabies vaccination or duration of protective RVNA values between horses < 20 and ≥ 20 years of age. Seven horses were poor responders to vaccination. Significant differences were identified between horses inferred to have been previously vaccinated and horses inferred to be naïve prior to the study. CONCLUSIONS AND CLINICAL RELEVANCE A rabies vaccination interval > 1 year may be appropriate for previously vaccinated horses but not for horses vaccinated only once. Additional research is required to confirm this finding and characterize the optimal primary dose series for rabies vaccination.