Background: Hyperglycemia is common and hard to control in surgical patients with diabetes. We retrospectively investigated short-term effects of continuous subcutaneous insulin infusion (CSII) in perioperative patients with diabetes. Patients and Methods: Perioperative patients with diabetes discharged between January 1, 2006 and January 1, 2012 were included. Glucose control and postoperative outcomes were compared between the patients using CSII or non-CSII insulin therapy. Results: We identified 108 pairs of patients matched by propensity and surgical category who were using CSII therapy (CSII group) or non-CSII insulin therapy (control group). The CSII group had significantly lower fasting glucose levels (on the first postoperative day, 9.06 -3.09 mmol/L vs. 11.05 -4.19 mmol/L; P = 0.003) and lower mean glucose levels (on the operation day, 9.93 -2.65 mmol/L vs. 12.05 -3.86 mmol/L; P = 0.001). The CSII group also had a lower incidence of fever (on the first postoperative day, 30.4% vs. 53.2%; P = 0.005). Furthermore, patients in the CSII group experienced significantly shorter postoperative intervals for suture removal (P = 0.02) and hospital discharge (P = 0.03). No significant difference in the total medical expenditure was observed between the two groups (P = 0.47). We also made a comparison between the 30 pairs of patients who were using CSII or multiple daily insulin injection therapy but observed no significant difference between these two therapies in glucose control or postoperative outcomes. Conclusions: Compared with non-CSII insulin therapy, even short-term implementation of CSII can improve the postoperative control of glucose, reduce the incidence of postoperative fever, and shorten the time for suture removal and discharge in surgical patients with diabetes.