Abstract: The in-situ recirculating rat perfused liver model in our lab was taken as an example. A set of methods were introduced for evaluating the viability of isolated liver perfusion. Firstly, a series of measurements including gross appearance of the liver, pressure of portal vein, pH of perfusate, concentration of potassium, levels of cytosolic enzymes and morphology were suggested to provide an adequate general assessment of viability. The liver should be a homogeneous, pinkish-brown color during perfusion; white spots caused by air emboli and red spots due to nonhomogeneous perfusion should be absent. The pressure of portal vein should not go beyond 8-14 mmHg or raise more than 0. 5 mmHg/30 min. The pH of perfusate should not fluctuate more than 0. 1/30 min and be 7. 4-7. 2 all the time. Sudden increase in perfusate potassium was forbidden. Significant changes of ALT(alanine aminotransferase)and AST(aspartate aminotransaminase)levels along with the perfusion time were not allowed. The perfused liver should show no other pathological changes except vacuolization after the whole perfusion. Secondly, further observation and examination of functional applicability were extremely necessary for different study purposes; obvious differences of potential markers and indexes between the group perfused with K-H perfusate only and that with medicine-contained perfusate were needed. In this paper, the isolated perfused liver model was used for studying drug-induced liver injury. Significant differences of perfusate ALT, AST and LDH(lactate dehydrogenase)levels, the damage markers, between the control and isoniazid groups were observed. Morphological analysis of the perfused liver in isoniazid group showed more serious vacuolization than the control group with slight necrosis, which means qualified functional viability for liver injury research.