BACKGROUND: Looked-after children (LAC) are at greater risk of teenage pregnancy than non-LAC, which is associated with adverse health and social consequences. Existing interventions have failed to reduce rates of teenage pregnancy in LAC. Peer mentoring is proposed as a means of addressing many of the factors associated with the increased risk of teenage pregnancy in this group. OBJECTIVE: To develop a peer mentoring intervention to reduce teenage pregnancy in LAC. DESIGN: Phase I and II randomised controlled trial of a peer mentoring intervention for LAC; scoping exercise and literature search; national surveys of social care professionals and LAC; and focus groups and interviews with social care professionals, mentors and mentees. SETTING: Three local authorities (LAs) in England. PARTICIPANTS: LAC aged 14-18 years (mentees/care as usual) and 19-25 years (mentors). INTERVENTION: Recruitment and training of mentors; randomisation and matching of mentors to mentees; and 1-year individual peer mentoring. PRIMARY OUTCOME: pregnancy in LAC aged 14-18 years. SECONDARY OUTCOMES: sexual attitudes, behaviour and knowledge; psychological health; help-seeking behaviour; locus of control; and attachment style. A health economic evaluation was also carried out. RESULTS: In total, 54% of target recruitment was reached for the exploratory trial and 13 out of 20 mentors (65%) and 19 out of 30 LAC aged 14-18 years (63%) (recruited during Phases I and II) were retained in the research. The training programme was acceptable and could be manualised and replicated. Recruitment and retention difficulties were attributed to systemic problems and LA lack of research infrastructure and lack of additional funding to support and sustain such an intervention. Mentees appeared to value the intervention but had difficulty in meeting weekly as required. Only one in four of the relationships continued for the full year. A future Phase III trial would require the intervention to be modified to include provision of group and individual peer mentoring; internal management of the project, with support from an external agency such as a charity or the voluntary sector; funds to cover LA research costs, including the appointment of a dedicated project co-ordinator; a reduction in the lower age for mentee recruitment and an increase in the mentor recruitment age to 21 years; and the introduction of a more formal recruitment and support structure for mentors. CONCLUSIONS: Given the problems identified and described in mounting this intervention, a new development phase followed by a small-scale exploratory trial incorporating these changes would be necessary before proceeding to a Phase III trial. FUNDING: This project was funded by the NIHR Health Technology Assessment programme and will be published in full in Health Technology Assessment; Vol. 19, No. 85. See the NIHR Journals Library website for further project information.