Objective: The objective of this study was to evaluate the haemorrhage risk of known and unknown cerebral arteriovenous malformations and their obstetric management. Methods: A retrospective review was performed and analysed 67 consecutive cases of arteriovenous malformation with pregnancy history. Results: Sixty-seven cases of arteriovenous malformation with pregnancy histories were identified. In 14 cases (20.9%) of arteriovenous malformation diagnosed before pregnancy, 11 cases were treated (10 embolisation and one surgery), there was no haemorrhage in 14 pregnancies, 14 healthy babies were delivered by caesarean section in 12 pregnancies (85.7%) and vaginal delivery in two pregnancies (14.3%). In 53 cases (89.1%) of arteriovenous malformation diagnosed during/ after pregnancy, there was one (1.6%) case of subarachnoid haemorrhage at 38 weeks' gestation in 64 pregnancies, 64 healthy babies were delivered by caesarean section in 11 pregnancies (17.2%) and vaginal delivery in 53 pregnancies (82.8%). This resulted in 1.6% (95% confidence interval 0-4.6%) haemorrhage rate per pregnancy in unknown arteriovenous malformations. Known arteriovenous malformation gravida was prone to caesarean section; however, vaginal delivery did not increase the haemorrhage risk in unknown arteriovenous malformation gravidas (1.8% vs. 0%, P = 1.000). Conclusion: Prior treatment for ruptured arteriovenous malformation could prevent its haemorrhage during pregnancy and the haemorrhage risk of unruptured arteriovenous malformation in pregnancies is low. Although known arteriovenous malformation gravida is prone to caesarean section, vaginal delivery seems not to increase the haemorrhage risk in unknown arteriovenous malformation gravidas.