Objective To compare the degree of glycemia variability(GV) and hypoglycemia incidence by dynamic glucose monitoring in elderly patients with type 2 diabetes mellitus (T2DM) who were treated with different premixed insulin. Methods A total of 64 patients (≥60 years ) with type 2 diabetes were included. The general data were matched, and all patients required insulin treatment. The patients were randomly divided into two groups, and treated with biphasic insulin aspart 30 or insulin lispro mix 25 respectively. Predicating the blood glucose concentrations within normal range, difference in GV and hypoglycemia incidence between the two groups were analyzed by dynamic glucose monitoring. Results When patients had normal level of blood glucose, there was no significant difference in total insulin between two groups [(38.0 ± 6.2) vs (40.0 ± 5.1) U/d, P＞0.05]. There was no statistical difference found in the mean blood glucose drift rate [(6.32 ± 1.43) vs (6.86 ± 1.51) mmol/L, P＞0.05] or glucose drift coefficient [(1.35 ± 0.22) vs (1.41 ± 0.13), P＞0.05]. Injection events in insulin lispro mix 25 hypoglycemic were less than that in insulin aspart 30 group, but the difference was not statistically significant (12% vs 8%, P＞0.05). Conclusion When the blood glucose was under good control, insulin aspart 30 and insulin lispro mix 25 have similar effects for elderly patients with type 2 diabetes, including the total amount of insulin, glucose rate of drift and the incidence of hypoglycemia.