Objective: To compare the efficacy and safety of a twice daily injection of insulin aspart (BIAsp) 30 and BIAsp50 in patients with type 2 diabetes mellitus (T2DM) poorly controlled with oral hypoglycemic agents (OHAs). Methods: In this 12 week prospective, randomized, parallel trial, a total of 80 T2DM patients, 59 +/- 10 years old with a disease duration of 9.3 +/- 6.6 years and HbA1c >7% despite large doses of metformin and sulfonylurea administration, were randomized to receive BIAsp30 (n = 40) or BIAsp50 (n = 40). The primary endpoint was a change in HbA1c at week 12. Results: The changes in HbA1c from baseline were -2.5% +/- 1.0% in the BIAsp50 group and -2.5% +/- 1.2% in the BIAsp30 group (p = .897). No difference was observed in the rate of HbA1c target achievement (<7.0%) between BIAsp50 (42.5%) and BIAsp30 (32.5%) (p = .495). The change in fasting plasma glucose (FPG) in the BIAsp50 group was lower than that in the BIAsp30 group (p < .001), while the change in two-hour postprandial blood glucose (2hPBG) was higher and blood glucose excursion was lower in the BIAsp50 group than that in the BIAsp30 group (p < .001, p < .001). A significant improvement in HbA1c was observed with BIAsp50 in subgroups with baseline blood glucose excursion >7.8 mmol/L or 2hPBG >17.6 mmol/L compared with BIAsp30. There were no differences in hypoglycemia or body weight between groups. Conclusions: Compared with BIAsp30, BIAsp50 showed greater efficacy in patients with baseline BG excursion >7.8 mmol/L or 2hPBG >17.6 mmol/L as well as good safety for hypoglycemia.