Fulminant hepatic failure is one of the most dramatic entities in clinical medicine, but experimental studies of its pathogenesis, evolution and treatment have, so far been limited by the lack of satisfactory animal models for testing new supportive treatment options. The variable aetiology, complex pathogenetic mechanisms and inconstant clinical evolution of human fulminant hepatic failure make it particularly difficult to establish an "ideal fulminant hepatic failure animal model" suitable for all studies: it is no longer mandatory to develop one single model serving all possible scientific needs, but the use of a specific model for a specific issue is more advisable. The currently available animal models of fulminant hepatic failure are the hepatotoxic, surgical and combined hepatotoxic and surgical models. From a general point of view, surgical models may be particularly appropriate for studying the consequences of hepatic necrosis on cerebral oedema. The anhepatic model is very useful for validating new supportive measures to bridge the period between the onset of fulminant hepatic failure and the time at which a suitable organ becomes available and, despite the many difficulties involved in their development, hepatotoxic models may still be useful for mimicking an acetaminophen overdose. The efficacy and reproducibility of a liver support system can be demonstrated by means of preclinical experimental models that mimic the specific application required in humans as closely as possible.