Background Southern and Eastern Africa is home to more than 2.1 million young people aged 15 to 24 years living with HIV. As compared with other age groups, this population group has poorer outcomes along the HIV care cascade. Young people living with HIV and the research team co-created the PEBRA (Peer Educator-Based Refill of ART) care model. In PEBRA, a peer educator (PE) delivered services as per regularly assessed patient preferences for medication pick-up, short message service (SMS) notifications, and psychosocial support. The cluster-randomized trial compared PEBRA model versus standard clinic care (no PE and ART refill done by nurses) in 3 districts in Lesotho. Methods and findings Individuals taking antiretroviral therapy (ART) aged 15 to 24 years at 20 clinics (clusters) were eligible. In the 10 clinics randomized to the intervention arm, participants were offered the PEBRA model, coordinated by a trained PE and supported by an eHealth application (PEBRApp). In the 10 control clusters, participants received standard nurse-coordinated care without any service coordination by a PE. The primary endpoint was 12-month viral suppression below 20 copies/mL. Analyses were intention-to-treat and adjusted for sex. From November 6, 2019 to February 4, 2020, we enrolled 307 individuals (150 intervention, 157 control; 218 [71%] female, median age 19 years [interquartile range, IQR, 17 to 22]). At 12 months, 99 of 150 (66%) participants in the intervention versus 95 of 157 (61%) participants in the control arm had viral suppression (adjusted odds ratio (OR) 1.27; 95% confidence interval [CI] [0.79 to 2.03]; p = 0.327); 4 of 150 (2.7%) versus 1 of 157 (0.6%) had died (adjusted OR 4.12; 95% CI [0.45 to 37.62]; p = 0.210); and 12 of 150 (8%) versus 23 of 157 (14.7%) had transferred out (adjusted OR 0.53; 95% CI [0.25 to 1.13]; p = 0.099). There were no significant differences between arms in other secondary outcomes. Twenty participants (11 in intervention and 9 in control) were lost to follow-up over the entire study period. The main limitation was that the data collectors in the control clusters were also young peers; however, they used a restricted version of the PEBRApp to collect data and thus were not able to provide the PEBRA model. The trial was prospectively registered on ClinicalTrials.gov (NCT03969030). Conclusions Preference-based peer-coordinated care for young people living with HIV, compared to nurse-based care only, did not lead to conclusive evidence for an effect on viral suppression. Trial registration clinicaltrials.gov, NCT03969030, https://clinicaltrials.gov/ct2/show/NCT03969030.