Rift Valley fever (RVF) is a zoonotic mosquito-borne virus disease of livestock and wild ruminants that has been identified as a risk for international spread. Typically, the disease occurs in geographically limited outbreaks associated with high rainfall events and can cause massive losses of livestock. It is unclear how RVF virus persists during inter-epidemic periods but cryptic cycling of the virus in wildlife populations may play a role. We investigated the role that free-living African buffalo (Syncerus caffer caffer) might play in inter-epidemic circulation of the virus and looked for geographic, age and sex patterns of Rift Valley fever virus (RVFV) infection in African buffalo. Buffalo serum samples were collected (n = 1615) in Kruger National Park (KNP), South Africa, during a period of 1996-2007 and tested for antibodies to RVF. We found that older animals were more likely to be seropositive for anti-RVFV antibody than younger animals, but sex was not correlated with the likelihood of being anti-RVFV antibody positive. We also found geographic variation within KNP; herds in the south were more likely to have acquired anti-RVFV antibody than herds farther north - which could be driven by host or vector ecology. In all years of the study between 1996 and 2007, we found young buffalo (under 2 years of age) that were seropositive for anti-RVFV antibody, with prevalence ranging between 0 and 27% each year, indicating probable circulation. In addition, we also conducted a 4-year longitudinal study on 227 initially RVFV seronegative buffalo to look for evidence of seroconversion outside known RVF outbreaks within our study period (2008-2012). In the longitudinal study, we found five individuals that seroconverted from anti-RVFV antibody negative to anti-RVFV antibody positive, outside of any detected outbreak. Overall, our results provide evidence of long-term undetected circulation of RVFV in the buffalo population.