After the first year of life, injury is the leading cause of mortality in children and adolescents in Australia and a major cause of morbidity and hospitalization. Studies from Europe and North American and more recently New Zealand have demonstrated that injury rates differ between different socio-economic groups, with a relatively consistent pattern of higher rates being found in lower socio-economic groups. To date, the pattern of child injury morbidity by socio-economic group has not been studied in Australia. Eighteen months of data from Brisbane and Melbourne collections of the National Injury Surveillance Unit data base were analysed and linked to census data from each postcode in the collecting area to calculate injury rates. Injury rates for each postcode were correlated with measures of disadvantage derived from the same census data. There was a consistent pattern of moderately strong statistically significant associations (Pearson correlation coefficients generally 0.30-0.60) found between measures of disadvantage and injury rate at the postcode level of aggregation. This association was present across cities, age groups, type of injury and for those hospitalized as well as those treated as outpatients. It is apparent that residence in a low income area is a significant predictor of child injury in Australia.