Objective: To compare different methods of gestational age (GA) measurement for ensuring effective zidovudine (ZDV) prophylaxis to prevent mother-to-child transmission of HIV. Methods: For 1398 HIV-infected women enrolled in a perinatal prevention trial, gestation durations were calculated based on GA estimated using ultrasound (US), date of last menstruation period (LMP), first fundal height (FH 1), and a specific algorithm was developed to provide a “reference” GA. The performance of each GA estimate was evaluated by the percentage of women who would have received ≥ 8 weeks ZDV, if prophylaxis was initiated at 28 weeks. Results: The performances of the algorithm, US, LMP, and FH 1 were 95.5%, 94.8%, 88.4%, and 83.7%, respectively. US and FH 1 were significantly better when estimated before and after 24 weeks, respectively. Conclusion: In situations where no US is available and LMP is not or imprecisely known, FH 1 can be used after 24 weeks to schedule ZDV initiation date.