A prospective randomised trial was performed to assess the diagnostic accuracy according to the time interval and the efficacy of endoscopic injection of hypertonic saline-epinephrine(HS- E)solution, consisting of 3.5% sodium chloride with 0.0045% epinephrine, for actively bleeding peptic ulcers, exposed vessel or blood clot on ulcer bed, or Mallory-Weiss tear. Over 24 month, emergency endoscopy in 180 patients admitted for upper gastrointestinal hemorrhage identified 51 patients with nonvariceal hemorrhage. The causes of bleeding were; gastric ulcer in 32; duodenal ulcer in 13; gastric cancer in 4; Mallory-Weiss tear in 2. With this method, the hemostatic effect was permanent in 40 cases(84.3%), temporary in 9 cases(11.8%), and failed in 2 cases(3.9%). By applying this method, the rate of emergency operation for patients with bleeding from the upper gastrointestinal tract was significantly reduced from 20.0% (8/40)to 3.9%(2/51)(p<0.05). Emergency endoscopy in acute UGI bleeding increases the accuracy of detection of actual bleeding sites(p<0.05), but if the endoscopic procedure was performed within 48 hours, the hemostatic rate was not affected(p<0.05). We concluded that hypertonic saline-epinephrine injection method could provide a simple maneuver with reasonable cost, high safety, and satisfactory hemostatic efficacy in the treatment of nonvariceal upper gastrointestinal bleeding.