To examine the effects of aerobic training (AT) and resistance training (RT) compared to standard care on glycemic control in South Asian Sri Lankan adults with Type 2 Diabetes Mellitus (T2DM). Randomized controlled trial (RCT) with parallel-group design recruited 86 sedentary Sri Lankans (aged 35–65 years) with T2DM into aerobic training (AT, n = 28), resistance training (RT, n = 28) and control (CN, n = 30) groups. Supervised progressive exercise training consisting of 75 min per session, 2 days per week for 12 weeks was conducted. The primary outcome was pre- and post-intervention absolute change in hemoglobin A1c (HBA1c). Secondary outcomes were serum lipids, liver enzymes, chronic inflammatory status, anthropometry, body composition and blood pressure. The absolute change in HbA1c of RT vs. CN was −0.08% (95% CI, 0.8% to −0.7%, p = 0.8) and AT vs. CN was −0.22% (95% CI, 0.95% to −0.5%). Subgroup analysis (n = 49) with a high baseline HbA1c (>7.5%), absolute reduction in HbA1c in exercise groups were statistically significant (RT vs. CN was −0.37%; 95% CI 1.3% to −0.6%, p = 0.04 and AT vs. CN was −0.57%; 95% CI 1.7% to −0.6%, p = 0.03). The effect sizes (total and subgroup HbA1c >7.5%) ranged from 0.7 to 1.0 in AT, 0.4 to 1.1 in RT compared to 0.35 to 0.6 for the CN. Secondary outcomes did not significantly differ among groups. Exercise training 2 days/week improved glycemic control in Sri Lankan adults with T2DM and the effects were significant in high baseline HbA1c (>7.5%) groups (RT > AT). •First adequately powered randomized controlled exercise trial conducted in Sri Lankans with type 2 diabetes mellitus.•Exercise improved glycemic control and resistance training was more effective in patients with poor glycemic control.•Two days of exercise, 150 min per week in total was more feasible to adhere which gave the desired glycemic benefits.