An evidence-based approach is followed, with grading of evidence by study design, to evaluate surgical treatment of a bleeding peptic ulcer. In contrast to endoscopic treatment, reports of surgical treatment are rare, with only five randomized trials having been identified. Epidemiologic studies have demonstrated that the incidence of emergency surgery has not changed despite major improvements with endoscopic treatment. There are no proven alternatives to emergency operation for massive bleeding uncontrollable by endoscopic procedures. There is some debate about surgery for rebleeding, but no randomized trial has assessed whether a second endoscopic treatment alone is preferable to surgery with or without repeated endoscopic treatment. Concerning the type of operative procedure, the existing body of evidence, including two randomized studies, indicates that patients are best served by a relatively aggressive surgical approach. Today the value of these studies is limited owing to prevention of ulcer recurrence by eradication and technical improvements of local procedures (e.g., arterial ligation). Early elective surgery was tested in two randomized studies and several uncontrolled series, which demonstrated that it may be beneficial in high risk groups and harmful in others. Indications for early elective surgery should be refined taking into account updated prognostic information and more effective endoscopic treatment. Because of a new understanding of ulcer disease the role of surgery has changed markedly within the last years, no longer aiming to cure the disease but primarily to stop the hemorrhage. Evidence, however, is not derived from properly randomized controlled trials but is based on theoretic arguments and knowledge from studies not primarily dealing with operative treatment.