Background/purpose Vaginal delivery, compared with Cesarean delivery, remains a less chosen mode of delivery for twin pregnancy. We studied the maternal and perinatal outcomes of twin pregnancy with different modes of delivery. Methods A retrospective study with data collected from a regional hospital, including vital twin pregnancies delivered at gestational age of 32 weeks and above. Medical charts were reviewed for prenatal conditions and postpartum outcomes. Results Ninety-eight pairs of twins were included and 44.9% were delivered via vaginal delivery. Women in the vaginal delivery group were significantly younger (32.5 ±4.3 years versus 34.8 ±4.6 years, p < 0.01), multiparous (34.1% versus 18.5%) and with more twins in vertex-vertex presentation (70.5% versus 33.3%) compared with women in the Cesarean delivery group. There were no differences between maternal postpartum complications and neonatal outcomes in both groups. The outcomes showed longer inter-twin delivery time interval (5.7 ± 5.6 versus 1.5 ± 0.9 min, p < 0.01), less estimated blood loss (198.7 ± 144.1 versus 763.2 ± 332.3 mL, p < 0.01), and shorter maternal hospital stay (3.0 ± 0.5 versus 5.7 ± 0.5 days, p< 0.01) in the vaginal delivery group. Twenty newborns had Apgar score below seven at birth. Logistic regression analysis revealed that low Apgar score was independently related to younger maternal age, maternal obstetric diseases and fetal non-vertex presentation. Gestational weeks and mode of delivery were not related to low Apgar score. Conclusion With careful case selection, vaginal delivery could be safely performed in twin pregnancies with less estimated blood loss and better recovery than Cesarean delivery.