One hundred and sixty five cases of uper gastrointestinal bleeding were examined by fiberoptic endoscopy at Busan Medical Center from July 19S4 to June 1987. The results obtained were as follows; 1) The sex incidence of upper gastrointestinal bleeding revealed a greater prevalence in males than in females with a ratio of 4 to 1 and the peak age group were in the 5th and 6th decades. 2) The causative disease which were found to protace upper gastrointestinal bleeding were in the order of frequency of peptic ulcer (43.0%), esophageal varices (21.8%), hemorrhagic gastritis(18.8%), stomach cancer (9.1%), Mallory-Weiss syndrome (3,6%) and marginal ulcer. The incidence of upper gastrointestinal bleeding of unknown origin was 3.1%. 3) On of cyamination the mode of upper gastrointestinal bleeding, the incidence of melenal only was noted in 36.4% of the cases, hemestasis only in 32.7% and hematemesis with melena in 30.9%. 4) The severity of upper gastrointestinal bleeding was classified by cammock's criteria as massive in 46.1%, moderate in 40.0% and mild in 13.9% of the cases. 5) The frequency of upper gastrointestinal bleeding at admission was first time bleeding in 41.8% of the cases, second time bleeding in 20.0% third time bleeding in 17.0% and over third time bleeding in 21.2%. 6) The success rate in identifying the bleeding site in relation to the time of endoscopy after upper gastrointestinal bleeding began was 69.4% within 24 hours, 54.3% in 24-48 hours, 17.9% In 48-72 hours and 7.7% after 7 days, The success rate fell with increasing time after upper gastrointestinal bleeding began, 7) The factors which influenced the outcome of upper gastrointestinal bleeding was age, severity of bleeding and episode of bleeding. The overall mortality rate was 9.7%