Our aim was to review systematically the literature on educational interventions to improve prescribing by medical students and junior doctors. MEDLINE, EMBASE, Educational Resource Information Center, British Education Index, PsycINFO, CINAHL, TIMELIT, Cochrane Trials Database and grey literature were searched. Inclusion criteria were: educational interventions to improve medical student and/or junior doctors' prescribing, in primary or secondary care settings, and published after 1990. After screening 3189 records, we retrieved 11 controlled and four 'before-and-after' trials. Ten controlled trials showed improvements in the scores of the intervention group on written scenarios or clinical examination stations, but one study in junior doctors showed no effect on real-life prescription errors. Only one intervention [the World Health Organization (WHO) Good Prescribing Guide, in six randomized trials] had been tested in a variety of international settings and across a range of students at different levels. All four 'before-and-after' trials reported significant improvements in written tests or clinical stations. However, most studies tested only small numbers of participants and were affected by a range of methodological flaws. There is only moderate evidence in the literature to inform medical schools about how to prepare medical students for the challenges of prescribing. The WHO Good Prescribing Guide is the only model that has been widely used and shown to improve prescribing. Although it is based on sound principles, there is a need for further development. Robust methods of assessment are required to show clearly whether particular teaching interventions are successful.