Patient-controlled epidural analgesia is widely used to control postoperative pain following major intra-abdominal surgeries. However, determining the optimal infusion dose that can produce effective analgesia while reducing side effects remains a task to be solved. Postoperative pain and adverse effects between variable-rate feedback infusion (VFIM group, n = 36) and conventional fixed-rate basal infusion (CFIM group, n = 36) of fentanyl/ropivacaine-based patient-controlled epidural analgesia were evaluated. In the CFIM group, the basal infusion rate was fixed (5 mL/h), whereas, in the VFIM group, the basal infusion rate was increased by 0.5 mL/h each time a bolus dose was administered and decreased by 0.3 mL/h when a bolus dose was not administered for 2 h. Patients in the VFIM group experienced significantly less pain at one to six hours after surgery than those in the CFIM group. Further, the number of patients who suffered from postoperative nausea was significantly lower in the VFIM group than in the CFIM group until six hours after surgery. The variable-rate feedback infusion mode of patient-controlled epidural analgesia may provide better analgesia accompanied with significantly less nausea in the early postoperative period than the conventional fixed-rate basal infusion mode following open gastrectomy.