Background and Aim: Reflux symptoms in patients with non-erosive reflux disease (NERD) cannot be easily controlled by treatment with proton pump inhibitors (PPI). The anti-inflammatory function of rebamipide may be effective for protecting the esophageal mucosa. This prospective randomized multicenter placebo-controlled study was performed to clarify the efficacy of rebamipide for NERD patients whose reflux symptoms were refractory to PPI treatment. Methods: One hundred forty-nine patients were enrolled on the basis of a QUEST score of over 6 and absence of endoscopically proven esophageal mucosal breaks. All the patients were initially administered 15 mg of lansoprazole for 4 weeks, and the symptoms were then assessed using QUEST and GSRS. PPI-refractory patients were randomly assigned to administration of rebamipide or placebo t.i.d. for 4 weeks. Results: Three of the 149 patients were lost to follow-up, and 60 among the remaining 146 patients were found to be PPI-refractory. Among these PPI-refractory patients, 31 were randomly assigned to a rebamipide group and 29 to a placebo group. At the end of drug administration, the QUEST and GSRS scores did not differ between the rebamipide and placebo groups, although a significantly higher proportion of patients in the rebamipide group showed amelioration of abdominal pain and diarrhea. Conclusion: Administration of rebamipide cannot effectively control reflux symptoms in NERD patients whose symptoms are refractory to PPI therapy.