Background Infant simulator-based programmes, which aim to prevent teenage pregnancy, are used in high-income as well as low-income and middle-income countries but, despite growing popularity, no published evidence exists of their long-term effect. The aim of this trial was to investigate the effect of such a programme, the Virtual Infant Parenting (VIP) programme, on pregnancy outcomes of birth and induced abortion in Australia. Methods In this school-based pragmatic cluster randomised controlled trial, eligible schools in Perth, Western Australia, were enrolled and randomised 1:1 to the intervention and control groups. Randomisation using a table of random numbers without blocking, stratification, or matching was done by a researcher who was masked to the identity of the schools. Between 2003 and 2006, the VIP programme was administered to girls aged 13–15 years in the intervention schools, while girls of the same age in the control schools received the standard health education curriculum. Participants were followed until they reached 20 years of age via data linkage to hospital medical and abortion clinic records. The primary endpoint was the occurrence of pregnancy during the teenage years. Binomial and Cox proportional hazards regression was used to test for differences in pregnancy rates between study groups. This study is registered as an international randomised controlled trial, number ISRCTN24952438. Findings 57 (86%) of 66 eligible schools were enrolled into the trial and randomly assigned 1:1 to the intervention (28 schools) or the control group (29 schools). Then, between Feb 1, 2003, and May 31, 2006, 1267 girls in the intervention schools received the VIP programme while 1567 girls in the control schools received the standard health education curriculum. Compared with girls in the control group, a higher proportion of girls in the intervention group recorded at least one birth (97 [8%] of 1267 in the intervention group vs 67 [4%] of 1567 in the control group) or at least one abortion as the first pregnancy event (113 [9%] vs 101 [6%]). After adjustment for potential confounders, the intervention group had a higher overall pregnancy risk than the control group (relative risk 1·36 [95% CI 1·10–1·67], p=0·003). Similar results were obtained with the use of proportional hazard models (hazard ratio 1·35 [95% CI 1·10–1·67], p=0·016). Interpretation The infant simulator-based VIP programme did not achieve its aim of reducing teenage pregnancy. Girls in the intervention group were more likely to experience a birth or an induced abortion than those in the control group before they reached 20 years of age. Funding Western Australian Health Promotion Foundation (Healthway), Lotteries WA, the Western Australian Department of Education and Training, and the Western Australian Department of Health.