Esophageal perforations are surgical emergencies associated with high morbidity and mortality rates. No single strategy has been sufficient to deal with the majority of situations. We aim to postulate a therapeutic algorithm for this complication based on 20 years of experience and also on data from published literature. We performed a retrospective clinical review of 44 patients treated for esophageal perforation at our hospital between January 1989 and May 2008. We reviewed the characteristics of these patients, including age, gender, accompanying diseases, etiology of perforation, diagnosis, location, time interval between perforation and diagnosis, treatment of the perforation, morbidity, hospital mortality, and duration of hospitalization. Perforation occurred in the cervical esophagus in 14 patients (31.8%), thoracic esophagus in 18 patients (40.9%), and abdominal esophagus in 12 patients (27.3%). Management of the esophageal perforation included primary closure in 23 patients (52.3%), resection in 7 patients (15.9%), and nonsurgical therapy in 14 patients (31.8%). In the surgically treated group, the mortality rate was 3 of 30 patients (10%), and 2 of 14 patients (14.3%) in the conservatively managed group. Four of the 14 nonsurgical patients were inserted with covered self-expandable stents. The specific treatment of an esophageal perforation should be selected according to each individual patient. To date, the most effective treatment would appear to be operative management. With improvements in endoscopic procedures, the morbidity and mortality rates of esophageal perforations are significantly decreased. We suggest that minimally invasive techniques for the repair of esophageal perforations will be very important in the future treatment of this condition.