Objective: The prenatal diagnosis of cardiac defects can potentially reduce postnatal morbidity and mortality. We wanted to evaluate prenatal cardiac diagnosis accuracy in a population referred for echocardiography. Methods: Single centre retrospective study of echocardiography referrals between April 1999 and December 2008. We compared the prenatal and postnatal cardiac diagnoses, the modified Aristotle and Wald scores. The final diagnosis Wald score was used to evaluate benefit. Results: Six hundred fetuses were included. Diagnoses included: normal heart (312, 52%); congenital heart defect (CHD) (231, 38.5%); primary arrhythmia (39, 6.5%); or cardiomyopathy, myocarditis or cardiac tumor (18, 3%). The prenatal and postnatal Aristotle and Wald scores correlated in 81% and 86%, respectively, each with significant differences in 22 cases. Four significant CHDs were misdiagnosed, the surgical prediction was incorrect in 7 and 13 false positive diagnoses of aortic coarctation were made. In 76% (455/600) fetuses prenatal diagnosis was considered beneficial. The average CHD Aristotle score was 9.5±5.0. In babies with CHDs and normal karyotype the score was either 6.5±5.0, 12.9±3.1 or 13.2±2.9, in survivors, cases of postnatal demise and cases of pregnancy termination, respectively. Conclusion: Prenatal diagnosis was accurate and the counselling appropriate in most cases; however, a few errors were made. The diagnosis of aortic coarctation remains challenging. © 2012 John Wiley & Sons, Ltd.