The majority of research on the benefits of physical activity to counter the occurrence of diabetes has employed continuous aerobic exercise, often performed at a low or moderate intensity. Recent research has identified intermittent exercise, as performed during team sports, as a promising alternative for those otherwise not inclined towards conventional aerobic exercise training, particularly given the associated greater enjoyment, motivation and social benefits. However, there is limited evidence for the efficacy of intermittent exercise for improving factors normally associated with the development of diabetes. In the present study, we investigated whether small-sided games (SSG) could be an effective alternative to traditional continuous exercise training. Thirty middle-aged, sedentary, men (48.6±6.6 y) were randomly assigned to a cycle ergometry (n=11), SSG (n=10), or control group (n=9). Participants in both exercise groups trained 3 d/wk for 8 wk, while control participants maintained normal activity and dietary patterns. All participants completed pre- and post-intervention testing, including: dual-energy x-ray absorptiometry scan, incremental exercise test, resting oral glucose tolerance test (OGTT) and a resting muscle biopsy. Plasma glucose and insulin concentrations were measured in a fasted state and throughout the 2 h OGTT. Total content of proteins associated with mitochondrial biogenesis, and glucose and insulin signaling, were analysed via western blotting. Two-way RM ANOVA with post hoc and effect sizes for difference between groups. Paired sample t-tests for within-group significance. Peak oxygen consumption improved by 19% in both exercise groups (P<0.05). In response to the OGTT, the control group exhibited no changes, while both the cycling and SSG groups showed a decline in the area under the curve (AUC) for glucose, and only SSG showed a decline in the AUC for insulin (P<0.05). There was a trend for an increase in PGC-1α (effect size (ES) ± confidence interval: 0.49±0.62), p53 (ES 0.47±0.55) and GLUT4 (ES 0.57±0.65) within the SSG group, and for mitochondrial complex I (P<0.05; ES 1.0±0.67) and IV (P<0.05; ES 1.1±0.77) within the cycling group. Participants in both exercise groups decreased total body fat mass (%) and intra-abdominal fat mass (kg; P<0.05). Collectively, SSG training in sedentary middle-aged populations was as effective as traditional cycle ergometry in reducing clinical risk factors normally associated with the development of diabetes. In addition, SSG training evoked skeletal muscle adaptations involving an increase in p53, PGC-1α and GLUT4 proteins, collectively important for mitochondrial biogenesis and glucose regulation. These changes show the potential relevance of intermittent exercise training via SSG as a chronic disease prevention strategy.