Objective Studies suggest that the prevalence of HIV is higher among long term marital/consensual relationships than in the unmarried. We assessed the risk of incident HIV infection by marital status in rural Rakai, Uganda. Design Longitudinal data from the Rakai Community Cohort Study (RCCS) between 1999 - 2011 Methods We estimated HIV incidence per 100 person years (py) in sexually active individuals aged 15-49 with a total of 44,179.6 person years (py) who were never married (females 2,929py, males 4,261py), currently married or in long-term consensual relationships (“currently married females 29,823py, males 21,299py) and previously married (females 3,563py, males 1,475). Poisson multivariable regression was used to estimate the unadjusted and adjusted incident rate ratios (IRRs) and 95% confidence intervals (95% CI) of HIV acquisition. Results HIV incidence among currently married persons was 0.93/100py, which was lower than the never married (1.51/100py) and previously married (2.85/100py). The risk of HIV acquisition was significantly lower in the currently married compared to the never married among women (Adj IRR=0.26, 95% CI: 0.16-0.42), but not men (Adj IRR=0.69, 95% CI: 0.31-1.52). HIV incidence was lower among first marriages (0.73/100py) compared to second or higher order marriages (1.38/100py). Multiple sex partners significantly increased the risk of HIV acquisition in both women (Adj IRR=2.53, 95% CI: 1.6, 3.97) and men (Adj IRR=1.77, 95% CI: 1.20-2.60). Conclusion Current marriage especially first order marriage was associated with reduced risk of HIV acquisition in women, but not in men, and multiple sex partnerships increased HIV risk for both sexes.