Objective To establish a mouse model of colon cancer liver metastasis with high metastasis rate,simple operation and reliable outcome, and to serve the studies on prevention and treatment of colon cancer metastasis.Methods Fifteen BALB/ c nude mice were divided into 3 groups of 5 mice each (A, B, and C), and group D consistingof 5 wild type BALB/ c mice. We established four colon cancer metastasis models: spleen planted, spleen preserving, andsplenectomy method using 0. 2 mL of HCT116 or CT26 cell suspension with a cell concentration of 2. 5 × 10⁷/ mL,respectively. We compared the success rates of modeling in the four groups as well as intra-abdominal metastasis, and livermetastasis size and number. Results The success rate of the group A was 100% (5/5). The number of liver metastases inthe group A was low, dispersed, and located in the right lobe of the liver. The mean survival time was (26. 6 + 3. 4) days.The success rate of Group B was 40% (2/5). Metastatic tumors in the group B were scattered on the surface of the liver,and the volume was larger than the group A. The mean survival time was (36. 8 ±4. 2) days. The success rate of the groupC was 100% (5/5). The number of liver metastases in the group C was higher than in the other groups, and multiplemetastatic tumors fused into a mass occupying the right lobe of the liver. The mean survival time was (20. 2 ±2. 6) days.No metastasis was found in the group D. Peritoneal metastasis occurred in the three groups of nude mice (n =2 in Group A,n =3 in Group C) without heart, lung, brain, or kidney metastasis. The histopathological examination of liver metastasesconformed to the characteristics of adenocarcinoma. Conclusions The spleen preserving method shows the highest rate ofmodel establishment, and effectively simulates the route and process of human colon cancer cells travelling to the liver by the blood circulation.