STUDY OBJECTIVE: This study investigates the risk and mediating mechanisms of unintentional and violent injuries in pre-school children of teenage mothers. DESIGN: Cohort study based on Swedish national registers. Cox analyses of proportional hazard were used to estimate the relative risk of hospital admission and death attributable to injuries in analyses of data from national registers. PARTICIPANTS: The study population was a national cohort of 800,192 children born in Sweden during 1987-93 who were followed up prospectively from birth to their 7th birthday. MAIN RESULTS: Children of teenage mothers had higher relative risks (RRs) of hospital admissions for violent as well as unintentional injuries; age adjusted RRs of 2.7 (95% CI 1.2 to 6.1) and 1.6 (1.4 to 1.8), respectively, for children of mothers under 18 years of age and 2.5 (1.6 to 3.8) and 1.5 (1.4 to 1.6) of mothers aged 18-19 are compared with those with mothers aged at least 32 at the birth of the child. When the models were adjusted to socioeconomic variables and indicators of parental substance misuse and psychiatric illness the risk decreased slightly but remained well above that of children with older mothers. In addition, children of teenage mothers had an increased risk of death attributable to violent injuries (RR 6.7 (2.6 to 16.0), as well as to unintentional injuries (RR 3.5 (2.0 to 6.1). CONCLUSIONS: Maternal age is an important determinant of injuries in pre-school children in Sweden and the children of teenage mothers are at particular risk. Young parents should be given priority in injury prevention programmes.