Background: Endoscopy is the most accurate method for diagnosing the source of upper gastrointestinal bleeding. This study was aimed to evaluate the correlation between the timing of elective endoscopy and the length of hospital stay, the amount of transfusion given and incidence of recurrent bleeding or patient mortality. Method: A retrospective study was conducted in all patients with non-variceal upper gastrointestinal bleeding who had experienced elective endoscopy at Cipto Mangunkusumo Hospital between January 2007 and August 2008. Identification of clinical risk using clinical Rockall score was performed at the emergency room. Persistent bleeding, recurrent bleeding, surgical treatment and death were the outcome variables. Statistical analysis was performed using Chi-square/fisher exact test and linear regression. Results: There were 40 eligible cases with mean age of 53 ± 13 years; the greatest occurrence was at the age group of 50-59 years (12%), male (52.5%) and those who had clinical symptom of melena (52.5%). Twenty seven (67.5%) patients had Rockall score of 1-3 points and 13 (32.5%) had 4-6 points. There was only one patient who had adherent clots (Forrest grade II B). Endoscopy results revealed that the most common cause of bleeding was gastric ulcer, which occurred in 12 (30%) patients. There was no correlation between the timing of endoscopic procedures and outcome variable; however the length of hospital stay had a significant correlation with timing of endoscopic procedures. Conclusion: Elective endoscopy does not affect the variables of mortality and recurrent bleeding; however, it affects the length of hospital stay. Further prospective studies are required to find causal relation between them.