Background Esophagogastroduodenoscopy (EGD) is currently considered the first-line diagnostic procedure of choice for upper gastrointestinal bleeding (UGIB); however, the etiology of bleeding remains unknown in a subset of patients. This study aimed to evaluate the diagnostic yield of EGD in UGIB in pediatrics and determine the clinical predictors for positive endoscopic diagnosis. Methods A cross-sectional study was conducted at the pediatrics endoscopy unit, Ain Shams University, Cairo, Egypt, where 100 children were included. They were referred for EGD due to overt UGIB in the form of hematemesis and/or melena. Full medical history, thorough physical examination, laboratory investigations, and endoscopic and histopathologic findings were documented. Results Forty-seven males and 54 females were included. Their ages ranged from 3 months to 15 years, with a median age of 4 years. Sixty-five percent presented with hematemesis only, 7% presented with melena only, and 28% presented with hematemesis and melena. An endoscopic diagnosis could be reached in 62% of cases, with Helicobacter pylori (H. pylori) gastritis (23%) and reflux esophagitis (11%) as the most common endoscopic diagnoses, with the former being the most common in children above 4 years and the latter for younger ones. Other diagnoses included non-specific gastritis (8%) and esophageal varices (4%). Presentation with melena only was a negative predictor to reach a diagnosis by EGD, while splenomegaly and thrombocytopenia were independent predictors of variceal bleeding. Conclusion EGD is the investigation of choice in children suffering from hematemesis especially in older age groups. Clinical and laboratory parameters might help in the prediction of the underlying etiology.