The objective of the study was to develop and evaluate an effective whole-community approach to identifying a diverse group of influential young people to effectively diffuse health promotion messages among their peers. A peer nomination questionnaire, developed through extensive piloting work, was completed by 10 730 Year 8 students (aged 12-13 years) in 59 schools (30 intervention, 29 control) as part of a cluster randomized controlled trial. Influential students identified in 30 intervention schools were trained to disseminate smoke-free health promotion messages through informal contacts with peers. This approach successfully identified, recruited and retained a diverse group of students, broadly representative of their year group, to undertake the role of 'peer supporter'. Although students and staff expressed doubts about the suitability of some young people recruited as peer supporters, the intervention achieved a 22% reduction in the odds of being a regular smoker in intervention compared with control schools [odds ratio 0.78 (95% CI 0.64-0.96)]. Carefully designed and developed peer-led interventions have potential for delivering effective smoking prevention among adolescents. Paying close attention to the way in which peer educators are identified, and involving young people themselves in this process, may be the key to increasing the effectiveness of peer education.