BACKGROUND: A population-based study on the prevalence of cerebral palsy has not been previously carried out in China. The purpose of the present paper was to determine the overall and birthweight-specific prevalence of cerebral palsy. METHODS: A cross-sectional study of cerebral palsy was carried out among 388192 children aged <7 years in seven cities of Jiangsu province in China from May to July 1997. Information about birthweight was obtained from routine health care records. Doctors from township and city hospitals examined all eligible children and doctors at city level finally diagnosed all cases. All the doctors involved had taken part in a training programme held by Beijing Medical University in April 1997. RESULTS: The total prevalence of cerebral palsy was 1.6 per 1000 children and the birthweight-adjusted prevalence 2.8 per 1000 children (using Australia's neonatal survivors 1994 as a standard population). The overall neonatal mortality rate was 6.8 per 1000 live births, being highest (581.1 per 1000) in the 1000-1499 g birthweight group. The birthweight-specific cerebral palsy prevalence ranged from 0.8 per 1000 children in children weighing 3750-3999 g to 67.3 in children weighing 1500-1749 g. Children weighing 3500-3999 g at birth were at the lowest risk of cerebral palsy. In a given low birthweight group the prevalence of cerebral palsy in China was higher than that in developed countries even though this study was unable to include those who died at risk of, or with cerebral palsy. However, the prevalence of cerebral palsy at normal birthweight was almost the same as that in developed countries. In all, about 2% of all children were of low birthweight (<2500 g), with those weighing <1500 g accounting for about 0.02%. Children weighing <2500 g at birth contributed 24% of all cerebral palsy cases with 99% in the group 1500-2499 g. CONCLUSIONS: The prevalence of cerebral palsy for children aged <7 years is 1.6 per 1000 children. It is estimated that there are 310000 children with cerebral palsy in China and as the survival of low birthweight infants improves the prevalence of cerebral palsy will rise. Survival of low birthweight infants is lower in China than in developed countries and our findings suggest the survival quality of these Chinese children needs to be improved and that intrapartum and neonatal antecedents might play an important role in the aetiology of cerebral palsy compared to developed countries.