We evaluated the influence of short-term treatment with exenatide twice daily or liraglutide once daily on daily blood glucose fluctuations in 40 patients with type 2 diabetes inadequately controlled by sulfonylureas. The patients in a multicenter, open-label trial were randomly assigned to receive add-on exenatide (10 μg/day, n = 21) or add-on liraglutide (0.3-0.9 mg/day, n = 19), and underwent 24-hour continuous subcutaneous glucose monitoring. There was no significant between-group difference in glucose fluctuations during the day, as assessed by calculating mean amplitude of glycemic excursion (MAGE) and standard deviation (SD). However, the mean blood glucose levels at 3 hours after breakfast and dinner were significantly lower in the exenatide group than the liraglutide group (breakfast: 127.3 ± 24.1 vs. 153.4 ± 28.7 mg/dL; p = 0.006, dinner: 108.7 ± 17.3 vs. 141.9 ± 24.2 mg/dL; p < 0.001). In contrast, mean blood glucose levels and their SD were significantly lower between 0000 h and 0600 h in the liraglutide group than the exenatide group (average glucose: 126.9 ± 27.1 vs. 107.1 ± 24.0 mg/dL; p = 0.029, SD: 15.2 ± 10.5 vs. 8.7 ± 3.8; p = 0.020). Both groups had similar glucose fluctuations despite differences in 24-hour blood glucose profiles. Therefore, each of these agents may have advantages or disadvantages and should be selected according to the blood glucose profile of the patient.