Objective To summarize the key points of animal modeling of osteosarcoma, and to provide reference and suggestions for improving the modeling methods and evaluation indexes.Methods The database was established by searching the relevant literature on the animal model of osteosarcoma in CNKI, Wanfang database and PubMed. The species of experimental animals, gender, modeling methods, types of cancer cell lines and detection indicators were summarized, and the database was established for statistical analysis. Results A total of 284 literatures were included. Statistical analysis found that BALB / c-nu / nu nude mice were preferred in osteosarcoma model animals(227 cases，75.17%), followed by SD rats (20 cases, 6.62%). Subcutaneous cell fluid transplantation in the back (66 cases, 21.85%), subcutaneous cell fluid transplantation in the axils (55 cases, 18.21%) or in situ cell fluid transplantation (51 cases, 16.89%) were used as modeling methods. Human MG-63 cells (100 cases, 33.11%) and mouse UMR-106 cells (39 cases, 12.91%) were selected as the cancer cell lines.The most detected indexes were tumor tissue apparent index (238 cases, 83.80%), tumor tissue HE staining (129 cases, 45.42%), Animal apparent indicator (94 cases, 33.10%)，tumor tissue immunohistochemistry (89 cases, 31.34%), etc. Conclusions At present, osteosarcoma BALB / c-nu / nu nude mice aged 4 to 6 weeks are used as experimental animals, and human MG-63 cell heterotopic transplantation (back and axillary transplantation) is used to establish the animal model, and the detection indexes of osteosarcoma are comprehensively evaluated by animal apparent index, tumor apparent index and tumor histopathology. It is suggested to select serum biochemical index, apparent index of tumor tissue as well, HE staining of tumor tissue and immunohistochemistry of tumor tissue to evaluate the model. However, there is still a lack of animal model preparation and evaluation criteria with high clinical consistency. In this paper, the advantages and disadvantages are summarized through literature mining and data analysis, in order to provide reference for the establishment of a good OS model and better application to OS mechanism research and new drug development.