Abstract Purpose This study examines the association between mental health problems in adolescence and general practice (GP) costs during adulthood up to age 50 in the UK. Methods We conducted secondary analyses of three British birth cohorts (individuals born in single weeks in 1946, 1958 and 1970). Data for the three cohorts were analysed separately. All respondents who participated in the cohort studies were included. Adolescent mental health status was assessed in each cohort using the Rutter scale (or, for one cohort, a forerunner of that scale) completed in interviews with parents and teachers when cohort members were aged around 16. Presence and severity of conduct and emotional problems were modelled as independent variables in two-part regression models in which the dependent variable was costs of GP services from data collection sweeps up to mid-adulthood. All analyses were adjusted for covariates (cognitive ability, mother's education, housing tenure, father's social class and childhood physical disability). Results Adolescent conduct and emotional problems, particularly when coexisting, were associated with relatively high GP costs in adulthood up to age 50. Associations were generally stronger in females than males. Conclusion Associations between adolescent mental health problems and annual GP cost were evident decades later, to age 50, suggesting that there could be significant future savings to healthcare budgets if rates of adolescent conduct and emotional problems could be reduced. Trial registration Not applicable.