Esophageal stricture occurs at a high rate after endoscopic submucosal dissection, especially after entire circumferential dissection, leading to poor quality of life. This retrospective cohort study evaluated the stricture rate in circumferential mucosal defect cases following modified steroid administration. We enrolled 22 consecutive patients who underwent entire circumferential endoscopic submucosal dissection for superficial esophageal cancer between April 2010 and April 2015 at our hospital. Until January 2013, a systemic steroid-prednisolone-was administered at 30 mg/day and tapered over 8 weeks in the original method group (original group). From February 2013, 30 mg of prednisolone was administered orally for 3 weeks and then the dose was reduced in 5 mg decrements every 3 weeks. This group was classified as the modified method group (modified group). We retrospectively compared the stricture rates between the two groups. The postoperative stricture rate was significantly lower in the modified group (36.4%; 4/11 patients) than in the original group (82%; 9/11 patients; P = 0.04). The mean number of endoscopic balloon dilatation procedures was significantly lower in the modified group (6.2 ± 11.3) than in the original group (19.4 ± 15.3; P = 0.023). Pneumonia and oral herpes infection, which are adverse events potentially associated with steroid administration, were observed in the original group. Candida esophagitis, arthritis, and steroid-related myopathy were observed in the modified group. This modified systemic steroid administration was effective for patients with entire circumferential mucosal defect. The safety of this method was also demonstrated.