Introduction. A smoking prevention program was developed to prepare children in elementary school for secondary school. This study assessed the effects on smoking in secondary school. Methods. In 2002, 121 schools in The Netherlands were randomly assigned to the intervention or control group. The intervention group received 3 lessons in 5th grade of elementary school and a second 3 lessons in 6th grade. The control group received "usual care". Students completed 5 questionnaires: before and after the lessons in 5th and 6th grade and in the first class of secondary school. At baseline, 3173 students completed the questionnaire; 57% completed all questionnaires. Results. The program had limited effect at the end of elementary school. One year later in secondary school significant effects on behavioral determinants and smoking were found. The intervention group had a higher intention not to smoke (beta=0.13, 95% confidence interval = 0.01-0.24) and started to smoke less often than the control group (odds ratio = 0.59,95% confidence interval = 0.35-0.99): smoking increased from 2.5% to 3.6% in the intervention group and from 3.2% to 6.5% in the control group. Girls showed the largest differences in smoking between intervention and control condition. Conclusions. A prevention program in elementary school seems to be effective in preventing smoking.