Background In women with a multiple pregnancy, spontaneous preterm delivery is the leading cause of perinatal morbidity and mortality. Interventions to reduce preterm birth in these women have not been successful. We assessed whether a cervical pessary could effectively prevent poor perinatal outcomes. Methods We undertook a multicentre, open-label randomised controlled trial in 40 hospitals in the Netherlands. We randomly assigned women with a multiple pregnancy between 12 and 20 weeks' gestation (1: 1) to pessary or control groups, using a web-based application with a computer-generated list with random block sizes of two to four, stratified by hospital. Participants and investigators were aware of group allocation. For women in the pessary group, a midwife or obstetrician inserted a cervical pessary between 16 and 20 weeks' gestation. Women in the control group did not receive the pessary, but otherwise received similar obstetrical care to those in the pessary group. The primary outcome was a composite of poor perinatal outcome: stillbirth, periventricular leucomalacia, severe respiratory distress syndrome, bronchopulmonary dysplasia, intraventricular haemorrhage, necrotising enterocolitis, proven sepsis, and neonatal death. Analyses were by modified intention to treat. This trial is registered in the Dutch trial registry, number NTR1858. Findings Between Sept 21, 2009, and March 9, 2012, 813 women underwent randomisation, of whom 808 were analysed (401 in the pessary group; 407 in the control group). At least one child of 53 women (13%) in the pessary group had poor perinatal outcome, compared with 55 (14%) in the control group (relative risk 0.98, 95% CI 0.69-1.39). Interpretation In unselected women with a multiple pregnancy, prophylactic use of a cervical pessary does not reduce poor perinatal outcome.