Women from a refugee background who resettle in high-income countries experience poorer perinatal outcomes in comparison to women from host countries. There is a paucity of research on how best to improve these outcomes.To report on the effectiveness of an Australian Refugee Midwifery Group Practice service on perinatal outcomes.We used inverse probability of treatment weighting to balance confounders and calculate treatment effect and compare maternal and neonatal outcomes for women from a refugee background who received Refugee Midwifery Group Practice care (n = 625), to those receiving standard care (n = 634) at a large tertiary hospital (1 January 2016-31 December 2019). Prespecified primary outcomes included: proportion of women attending ≥ 5 antenatal visits, preterm birth (37 weeks), spontaneous onset of labour, epidural analgesia in the first stage of labour, normal birth (term, spontaneous onset, vertex, spontaneous vaginal birth, no epidural, no episiotomy), and exclusively breast-feeding at discharge.Women who received Refugee Midwifery Group Practice care were more likely to have spontaneous onset of labour (adjusted odds ratio 2·20, 95% CI 1·71-2·82; p 0·0001), normal birth (1·55, 1·23-1·95; p 0·0001), and less likely to use epidural analgesia (0·67, 0·50-0·89; p = 0·0067) and have a preterm baby (0·60, 0·36-0·99; p = 0·047). There was no difference between groups in women attending ≥ 5 antenatal visits and exclusive breastfeeding at discharge from hospital.A Refugee Midwifery Group Practice is feasible and clinically effective.Similar services could potentially improve outcomes for women from a refugee background who resettle in high-income countries.