Objective To establish a model of primary liver cancer in mice using three method , and to compare their advantages and disadvantages. Methods Eighty male C57BL/6 mice aged 14 – 16 days were randomly divided into four groups: control, low-dose diethylnitrosamine (DEN), high-dose DEN, and DEN + carbon tetrachloride (CCl₄). Mice in the DEN+CCl₄ group were intraperitoneally injected with 2 mg/ kg DEN, and then two weeks later CCl₄(5 mL/ kg, 20%) was intraperitoneally injected twice a week for 16 weeks. The control group was not administered any treatment. Mice in the low-dose DEN group were injected intraperitoneally with 25 mg/ kg DEN, while mice in the high-dose DEN group were intraperitoneally injected with 40 mg/ kg DEN. Survival curves were generated for mice in each group. Twenty-four weeks after establishing the model, mice in each group were anesthetized and sacrificed, and the liver was dissected. The appearance of liver, incidence of tumors, number of tumors, and serum levels of alanine aminotransferase ( ALT), aspartate aminotransferase (AST), Golgi protein 73 ( GP73), alpha fetoprotein ( AFP) were measured. Results Compared with the control group, ALT ( P < 0. 001), AST ( P < 0. 001), GP73 ( P < 0. 001), AFP ( P < 0. 01), tumor incidence ( P < 0. 001), and tumor number in the three groups were significantly increased ( P < 0. 001). By 24 weeks after modeling, three mice in the low-dose DEN group had died, while six out of the 17 surviving mice developed liver cancer, resulting in an incidence of 35%. In the high-dose DEN group, six mice died and 12 out of the 14 surviving mice developed liver cancer, resulting in an incidence rate of 86%. In the DEN+CCl₄ group, four mice died and 9 out of the 16 surviving mice developed liver cancer, resulting in an incidence rate of 56%. All mice in the control group survived and no liver abnormalities were observed. Conclusions The liver cancer mouse model was successfully established in all experimental groups, but the high-dose DEN group exhibited obvious advantages with regard to successful establishment of a primary liver cancer model with high incidence and stability in a short period of time. Our result provide a new method for the establishment of a liver cancer model in mice.