To evaluate the diagnostic accuracy of fetal cerebral magnetic resonance imaging (MRI) on a large cohort and to compare pre- and postnatal MRI data. This prospective study included all cases referred to our unit for fetal cerebral MRI examination between June 2006 and December 2009 and which underwent at least one postnatal MRI examination. Cases in which there was termination of pregnancy, fetal death or stillbirth were excluded. The pre- and postnatal diagnoses established by MRI were compared and divided into five subgroups: same diagnosis on pre- and postnatal MRI (Group 1); same diagnosis but different appearance related to the natural course of the disease (Group 2); different diagnosis (related to limitations of fetal MRI) (Group 3); same diagnosis but with additional findings discovered on postnatal MRI examination (Group 4); or same diagnosis but different appearance related to the natural course of the disease (as in Group 2) and associated with additional findings discovered on postnatal MRI examination (Group 5). The prognostic impact of a possible disagreement between pre- and postnatal findings was evaluated. One hundred fetuses were included. Fetal MRI was performed at a mean gestational age of 33 (range, 24-39) weeks and postnatal MRI at a mean age of 3.5 months. There were 53 cases classified as Group 1, 32 in Group 2, four in Group 3, 10 in Group 4 and one in Group 5. Thus, in 15 cases (Groups 3-5), there were discrepancies between pre- and postnatal findings (mostly related to corpus callosum anatomy, cortical and migration disorders). The discrepancy was judged to have a prognostic impact in 9/15 cases. Two postnatal MRI examinations were performed in eight cases, in one of which the second examination showed subependymal heterotopia which were not detectable on the first examination. Pre- and postnatal MRI data showed good agreement in 85% of cases. There was disagreement with a prognostic impact in 9% of cases.