Universal eligibility for lifelong antiretroviral therapy (ART) for pregnant and breastfeeding women ('Option B+') has been widely adopted, but concerns remain. We tested the hypothesis that the change from CD4+-guided ART eligibility ('Option A'), to Option B+, would improve maternal ART uptake and retention. A stepped-wedge evaluation at 12 health facilities in eSwatini. Primary outcome was maternal retention: proportion of women attending clinic within 56 days of delivery (antenatal retention) and clinic attendance within 84 days of 6-months postpartum (postnatal retention). Generalized estimating equations examined impact of Option B+ vs. Option A. Between 19 August 2013 and 29 August 2014, 2347 HIV-positive women, 55% (n = 1296) Option A, 45%, (n = 1051) Option B+ were included. ART initiation was observed in 36% (n = 469) of Option A women vs. 94% (n = 983) under Option B+ (P < 0.001). Overall 39% (n = 912) were retained from first ANC visit through 6-months postpartum. Retention was higher under Option B+ (53%, n = 559) vs. Option A (24%, n = 353) with variation by site and study month. Adjusting for age, gestational age, previous HIV diagnosis, and CD4+, Option B+ women were significantly more likely to be retained antenatally (aRR 1.32; 95% CI 1.18-1.49; P < 0.001) and postnatally (aRR 2.11; 95% CI 1.79-2.49) compared with Option A. Restricted to women initiating ART, retention was lower under Option B+ (57%, n = 558) vs. Option A (66%, n = 309; aRR, 0.82; 95% CI 0.70-0.95; P < 0.0001). Compared with CD4+-guided ART eligibility, universal ART resulted in substantial increases in pregnant women initiating ART and retained in care through 6 months postpartum.