The cause and clinical course of the postoperative ARDS is, as of yet, not very well understood. The current study is a review of our experience on patients with ARDS after thoracotomy. Material and Method: Between Jan. 1996 to Aug. 2001, a retrospective analysis was conducted on 32 post-thoracotomy ARDS patients among 4018 patients receiving thoracotomy inclusive of thoracoscopic surgery. Result: The incidence of ARDS after pneumonectomy cases was 5.3%(13/245), 1.3% after lobectomy(9/ 710), and 4.4% after esophageal surgery(10/226). Of the 32 ARDS patients, 31 had malignant disease. The remaining 1 patient had aspergillosis. In the majority, the cause of ARDS was unknown. The average onset was on the 7.4th postoperative day. In 10 cases, the initial lesion was in the right lower lung field(31.2%), in the left lower lung field in 9(28.1%), and in both lower lung fields in 12(37.5%) cases. In all, the initial lesion was in the lower lung fields in 96.9% of the cases(31/32). There was a significant relationship between the development of ARDS and intraoperative I/O balance. The overall mortality rate was 65.6%(21/32). In the earlier period of the study(1996-Jun, 1998) the mortality rate was 100%, but in the latter period(July, 1998-Aug, 2001) it was significantly reduced to 47.6%: Conclusion: The current data showed a higher incidence of postoperative ARDS in patients with malignant disease and in those receiving extensive lymph node dissection with either lobectomy or pneumonectomy, and also in patients receiving esophageal surgery. In addition, introperative fluid overload was also associated with an increased incidence of ARDS. Treatment outcome could be improved with prone positioning and NO gas inhalation.