The aim of the study was to provide logical steps, alternative to the sequential chamber analysis, for diagnosing major congenital heart malformation prenatally. Between 1988 and 1995, of the 14,885 high-risk pregnancies screened at the Departments of Obstetrics and Paediatric Cardiology, University of Hong Kong, 92 fetuses with major congenital heart malformations were detected at 17-36 weeks' gestation (mean=24.6 weeks) by extended fetal echocardiographic screening. These were classified according to obvious abnormalities related to the partition of the heart or ventricular dimensions depicted by the four-chamber view. Subgrouping according to the presence of normal or abnormal ventriculo-arterial connection, with or without outflow tract anomalies, was further attempted. Five groups were identified: (1) septal defects with all four chambers identified (n=62, 67 per cent); (2) intact septation with dilatation of either the right or all four chambers of the heart (n=12, 13 per cent); (3) intact septation with hypoplasia of either one of the ventricles (n=11, 12 per cent); (4) a rudimentary interventricular septum giving an apparent three-chamber heart (n=6, 7 per cent); (5) miscellaneous causes (n=1, 1 per cent). Combined with the information related to the ventricular outflow tracts, the accuracy of arriving at a definitive diagnosis was 97 per cent, when compared with postnatal investigations, surgery and/or autopsies. This simple classification could provide logical steps to arrive at a definitive diagnosis for most major fetal heart malformations.