BACKGROUND: Truancy has been linked to risky sexual behaviours in teenagers. However, no studies in England have examined the association between truancy and teenage pregnancy, and the use of truancy as a marker of teenagers at risk of pregnancy. METHODS: Using logistic regression, we investigated the association between truancy at age 15 and the likelihood of teenage pregnancy by age 19 among 3837 female teenagers who participated in the Longitudinal Study of Young People of England. We calculated the areas under the ROC curves of four models to determine how useful truancy would be as a marker of future teenage pregnancy. RESULTS: Truancy showed a dose-response association with teenage pregnancy after adjusting for ethnicity, educational intentions at age 16, parental socioeconomic status and family composition ('several days at a time' versus 'none', odds ratio 3.48 95% confidence interval 1.90-6.36, P < 0.001). Inclusion of risk behaviours improved the accuracy of predictive models only marginally (area under the ROC curve 0.76 full model versus 0.71 sociodemographic characteristics only). CONCLUSIONS: Truancy is independently associated with teenage pregnancy among English adolescent girls. However, the discriminatory powers of models were low, suggesting that interventions addressing the whole population, rather than targeting high-risk individuals, might be more effective in reducing teenage pregnancy rates.