Abstract Introduction The objective of this study was to investigate the effect of adding exenatide to continuous subcutaneous insulin infusion (CSII) therapy on the precise insulin doses required by type 2 diabetic patients to maintain glycemic control. Methods This was a single-center, randomized, controlled, open-label trial. Uncontrolled T2D patients were recruited between March 2010 and November 2011 at Nanjing First Hospital, China. Subjects were randomly assigned (1:1) to either an exenatide add-on to CSII group or a CSII therapy only (i.e., control) group (n = 18, respectively) for 5 weeks. Patients were subjected to 3 days of continuous glucose monitoring (CGM) during the screening period and after therapy. The precise insulin doses, the times taken by the patients to achieve euglycemic control, and the mean amplitude of glycemic excursion (MAGE) at the endpoint were compared between the two groups. The primary endpoint was precise insulin dose differences between groups from baseline to the endpoint. Results A total of 36 subjects were admitted as inpatients. Patients in the exenatide add-on therapy group needed less insulin titration time to achieve glycemic control (3.67 ± 1.33 vs. 4.78 ± 1.00 days, P = 0.028) and significantly lower bolus insulin doses than the control group at the endpoint (total bolus, 0.13 ± 0.03 vs. 0.17 ± 0.04 U/kg, P = 0.02, breakfast bolus, 0.05 ± 0.01 vs. 0.06 ± 0.01 U/kg, P = 0.01, lunch bolus, 0.04 ± 0.01 vs. 0.06 ± 0.01 U/kg, P = 0.01, dinner bolus, 0.04 ± 0.01 vs. 0.05 ± 0.01 U/kg, P = 0.01, respectively). Moreover, the CGM data showed that patients in the exenatide add-on therapy group exhibited a significant reduction in MAGE as compared to the control group (2.96 ± 1.14 vs. 4.21 ± 1.39 mmol/L, P = 0.012). Conclusion Our data suggest that adding exenatide therapy to CSII therapy leads to an improvement in glycemic excursions and the use of smaller bolus insulin doses. Trial Registration Chinese Clinical Trial Registry identifier, ChiCTR-PPR-15007045.