Background: With proton pump inhibitors and current sophisticated endoscopic techniques, the number of patients requiring surgical intervention for upper gastrointestinal bleeding has decreased considerably while trans-arterial embolization is being used more often. There are few direct comparisons between the effectiveness of surgery and embolization. Methods: A retrospective study of patients from two Australian teaching hospitals who had surgery or trans-arterial embolization (n = 103) for severe upper gastrointestinal haemorrhage between 2004 and 2012 was carried out. Patient demographics, co-morbidities, disease pathology, length of stay, complications, and overall clinical outcome and mortality were compared. Results: There were 65 men and 38 women. The median age was 70 (range 36-95) years. Patients requiring emergency surgical intervention (n = 79) or trans-arterial embolization (n = 24) were compared. The rate of re-bleeding after embolization (42%) was significantly higher compared with the surgery group (19%) (P = 0.02). The requirement for further intervention (either surgery or embolization) was also higher in the embolization group (33%) compared with the surgery group (13%) (P = 0.03). There was no statistical difference in mortality between the embolization group (5/24, 20.8%) and the surgical group (13/79, 16.5%) (P = 0.75). Conclusion: Emergency surgery and embolization are required in 2.6% of patients with upper gastrointestinal bleeding. Both techniques have high mortalities reflecting the age, co-morbidities and severity of bleeding in this patient group.