Cerebral hypoxia has been proposed as a mechanism by which prenatal ethanol exposure causes fetal alcohol spectrum disorder (FASD) in children, but no study had tested this hypothesis using a chronic exposure model that mimicks a common human exposure pattern. Pregnant sheep were exposed to ethanol, 0.75 or 1.75 g kg−1 (to create blood ethanol concentrations of 85 and 185 mg dl−1, respectively), or saline 3 days per week in succession (a ‘binge drinking’ model) from gestational day (GD) 109 until GD 132. Fetuses were instrumented on GD 119–120 and studied on GD 132. The 1.75 g kg−1 dose resulted in a significant increase in fetal biventricular output (measured by radiolabelled microsphere technique) and heart rate, and a reduction of mean arterial pressure and total peripheral resistance at 1 h, the end of ethanol infusion. The arterial partial pressure of CO2 was increased, arterial pH was decreased and arterial partial pressure of O2 did not change. Fetal whole‐brain blood flow increased by 37% compared with the control group at 1 h, resulting in increased cerebral oxygen delivery. The elevation in brain blood flow was region specific, occurring preferentially in the ethanol‐sensitive cerebellum, increasing by 44% compared with the control group at 1 h. There were no changes in the lower dose group. Assessment of regional differences in the teratogenic effects of ethanol by stereological cell‐counting technique showed a reduced number of cerebellar Purkinje cells in response to the 1.75 g kg−1 dose compared with the control brains. However, no such differences in neuronal numbers were observed in the hippocampus or the olfactory bulb. We conclude that repeated exposure to moderate doses of ethanol during the third trimester alters fetal cerebral vascular function and increases blood flow in brain regions that are vulnerable to ethanol in the presence of acidaemia and hypercapnia, and in the absence of hypoxia.