To assess the effectiveness of non-surgical treatment for severe blunt liver injuries, we retrospectively examined the files of 147 patients (excluding cases with cardiopulmonary arrest on arrival). The patients were divided into three groups according to the date of admission: Group-A (1982-85), B (1986-90) and C (1991-95). The patients were further divided on the basis of the type of injury, according to the classification proposed by the Japanese Association for the Surgery of Trauma (JAST). The types of injuries reported here were of type II, IIIa or IIIb. There was no difference in the patients' background, injury severity score (ISS), amount of intra-peritoneal hemorrhage, and amount of blood transfusion among the three groups, A, B and C. Surgical treatment was carried out in all Group-A patients, in 81.8% of Group-B patients and in 31% of Group-C patients. The hospitalization period was shorter in the Group-C, and the frequency of fatal outcome was also lower in Group-C. Non-surgical treatment has been performed since 1988 in our center, and its frequency has been increasing since then. It was applied to 6 cases (18.2%) of Type II injuries in Group-B. In total, 69% of the cases in Group-C were non-surgically treated, including cases of Type IIIa and Type IIIb injuries. Non-surgical treatment prevented 8 out of 9 cases of severe liver injury from having fatal outcomes in 1995. Moreover, 5 of those cases were of the most serious type (Type IIIb). There were three fatalities in Group-C and none in Group-B. Out of all the non-surgically treated cases, injuries to the head were the cause of fatality. Any treatment of cases showing inferior vena cava damage was generally ineffective due to heavy bleeding, unless the bleeding was stopped by surgical intervention. Non-surgical treatment is suitable for blunt liver injuries, including severe liver injuries, provided that bleeding control and close observation are maintained.