Effective prevention of mother-to-child transmission benefits from early presentation to antenatal care (ANC). It is, however, unclear whether a previous HIV diagnosis results in earlier initiation of ANC. We estimated the probability of early ANC initiation among women with a previous HIV-positive diagnosis compared to those who first tested for HIV during ANC and explored determinants of early ANC among HIV-positive women. We conducted an analysis of a cross-sectional survey among 411 HIV-positive adult (>18 years) women who gave birth at midwife obstetrics units in Gauteng between October 2016 and May 2017. Predictors of early ANC (defined as initiating ANC before or at 14 weeks of gestation) were assessed by multivariate log-binomial regression model. Overall, 51% (210) were diagnosed during pregnancy with 89% (188) initiating antiretroviral therapy on the same day of diagnosis. There was no meaningful difference in the timing of ANC initiation between women with previous HIV diagnosis [adjusted risk ratio (aRR) = 1.2; 95% confidence interval (95% CI): 0.9-1.7] compared with those diagnosed during pregnancy. Early ANC was predicted by planned pregnancy [aRR = 1.3; 95% CI: 1.1-1.7], parity (>2 children) [aRR = 0.6; 95% CI: 0.2-0.9] compared to not having a child, and tuberculosis diagnosis [aRR = 2.9; 95% CI: 1.4-6.1]. Our results suggest the need for a targeted intervention among HIV-positive women by improving the quality, content and outreach of ANC services to enhance early ANC uptake, and minimize mother-to-child transmission risk.