To compare the effects of different approaches to establishing rat models of acute liver failure (ALF). Sixty-eight Sprague-Dawley rats were randomly divided into 3 groups for establishing ALF models using 3 different approaches, namely conventional hepatectomy for resecting 90% liver tissue as described by Higgins and Anderson, modified bloodless hepatectomy for resecting 90% liver tissue, and intraperitoneal injections of 700 mg/kg D-galactosamine (D-gal) and 5 µg/kg lipopolysaccharide (LPS). The mortality of the rats due to postoperative bleeding and survival rate at 7 days after the surgery were recorded. The levels of alanine aminotransferase (ALT), total bilimbin (Tbil), albumin (ALB), NH3, glucose (Glu) and prothrombin time (PT) were monitored, and histopathologies of the liver were examined at 24 and 72 h after the surgery. The mortality rate due to postoperative bleeding was higher in conventional hepatectomy group than in the modified surgical group (15% vs 0). The survival rate at 7 days was 25%, 0