Objective We aimed to improve an existing modeling method to establish an experimental model suitable for cancer research with short experimental period, low mortality, and modeling cost for metastasis. Methods Animals were injected intraperitoneally with 200 mg / kg N-diethylnitrosamine (DEN) at the beginning of the experiment. All animals in the first and second groups were given 40 or 80 ppm N-nitrosomorpholine (NMOR) via drinking water for 20 or 10 weeks. Animals in a third group were given 0. 8 mg / kg NMOR via gavage for 17 weeks. Animals in the second and third groups were fed for 7 weeks following the NMOR treatment. All animals in the first and third groups were subjected to two-thirds partial hepatectomy ( PH) to enhance cell proliferation activity at the end of the second week after DEN treatment. Results The first group had a liver tumor incidence of 10%, a survival rate of 100%, and a lung metastasis rate of 0% at 20 weeks. The second group had a liver tumor incidence of 35%, a survival rate of 76%, and lung metastasis rate of 0% at 17 weeks. The liver tumor incidence was 100% in the third group. The survival rate was 100% or 55% at 20 or 24 weeks. The lung metastasis rate was 55%. Conclusions Overall, these result suggest that modeling method such as intragastric administration of 0. 8 mg / kg NMOR for 17 weeks after intraperitoneal injection of 200 mg / kg DEN and a two- thirds PH can effectively establish an advanced liver tumor model with a high survival rate and morbidity in 20 weeks. In addition, metastases can appear at the later stages. This modeling method has several advantages, such as high incidence of hepatocellular carcinoma, high model survival rate, and late-stage metastasis. This could be an ideal model for research on the prevention and mechanism of hepatocellular carcinoma.