Introduction: The number of patients who performed pancreas division using stapler in distal pancreatectomy has been widely increasing. But, postoperative pancreatic fistula(POPF) still remains important problem. In this study, we analyzed our results retrospectively, and evaluate our new strategy to reduce POPF. Methods: 180 patients who underwent distal pancreatectomy were retrospectively analyzed. Grade B or C of the grading system of International Study Group of Pancreatic Fistula were considered to be POPF. Because we changed pancreatic division method from January 2005, we separate two periods in this study. The pancreatic stump was closed by hand-sewn suture or using stapler. The thickness of pancreas was measured in preoperative CT. Results: In first period, there was no significant difference in the incidence of POPF between hand-sewn group(17.0%) and stapler group(17.4%). In stapler group, the average pancreatic thickness in patients with POPF was 18.0mm and in patients without POPF was 13.5mm, respectively. There was significant difference. A 16mm cut-off for pancreatic thickness was established based on the receiver operating characteristic (ROC) curve. Therefore, we used stapler in case with under 16mm of pancreatic thickness and chose hand-sewn closure with over 16mm in second period. The incidence of POPF were 17.2% in first period and 10.9% in second period. The rate of POPF tended to decrease. The incidence of POPF with stapler were 17.4% in first period and 5.4% in second period. There was significant difference. Conclusions: To limit using stapler with a pancreatic thickness under 16mm, we could reduce the incidence of POPF.