Abstract Introduction Osteoporosis is a serious skeletal disease and as there is currently no cure, there is a large emphasis on its prevention, including the optimisation of peak bone mass. There is increasing evidence that regular weight-bearing exercise is an effective strategy for enhancing bone status during growth. This systematic review evaluates randomised and non-randomised controlled trials to date, on the effects of exercise on bone mineral accrual in children and adolescents. Methods An online search of Medline and the Cochrane database enabled the identification of studies. Those that met the inclusion criteria were included in the review and graded according to risk for bias. Results Twenty-two trials were reviewed. Nine were conducted in prepubertal children (Tanner I), 8 in early pubertal (Tanner II–III) and 5 in pubertal (Tanner IV–V). Sample sizes ranged from n = 10 to 65 per group. Exercise interventions included games, dance, resistance training and jumping exercises, ranging in duration from 3 to 48 months. Approximately half of the trials ( n = 10) included ground reaction force (GRF) data (2 to 9 times body weight). All trials in early pubertal children, 6 in pre pubertal and 2 in pubertal children, reported positive effects of exercise on bone ( P < 0.05). Mean increases in bone parameters over 6 months were 0.9–4.9% in prepubertal, 1.1–5.5% in early pubertal and 0.3–1.9% in pubertal exercisers compared to controls ( P < 0.05). Conclusions Although weight-bearing exercise appears to enhance bone mineral accrual in children, particularly during early puberty; it remains unclear as to what constitutes the optimal exercise programme. Many studies to date have a high risk for bias and only a few have a low risk. Major limitations concerned selection procedures, compliance rates and control of variables. More well designed and controlled investigations are needed. Furthermore, the specific exercise intervention that will provide the optimal stimulus for peak bone mineral accretion is unclear. Future quantitative, dose–response studies using larger sample sizes and interventions that vary in GRF and frequency may characterise the most and least effective exercise programmes for bone mineral accrual in this population. In addition, the measurement of bone quality parameters and volumetric BMD would provide a greater insight into the mechanisms implicated in the adaptation of bone to exercise.