We assessed whether HIV-1 viral load affects the likelihood of live birth among HIV-positive women in a nested case-control study of HIV-positive women from a community cohort in Rakai District, Uganda. Cases were women who had a live birth (n = 270), and controls were sexually active women who did not use contraception and did not become pregnant during follow-up (n = 263). In women with a live birth and non-pregnant controls, median HIV viral loads were 4.12 log(10) copies/mL and 4.41 log(10) copies/mL, respectively (P = 0.001). A non-linear association was observed, and a segmented linear regression with spline knot at 4.5 log(10) copies/mL was fit. We observed a decline in the log (adjusted odds ratio [adj. OR])= -0.08 (95% confidence interval [CI]: -0.36, 0.20) between 3.0 and 4.49 log(10) viral load and -0.92 (95% CI: -1.21, -0.63) between 4.5 and 6.5 log(10) viral load. The two reductions differed significantly from one another (P < 0.001). Each increase in log(10) viral load after 4.5 log(10) resulted in an adj. OR of live birth which was 12% of the previous viral load category. Our data suggest that there may be considerable differences in the ability to produce a live birth among HIV-positive women with high viral loads.