Abstract: OBJECTIVE To use data mining technology in the process of animal modeling to create a computer simulation model, and show the value of data mining technology. METHODS Thirty male SD rats were randomly divided into control group and model group, administered with intradermal injection of 0.25 mL normal saline and 0.25 mL complete Freund’s adjuvant, respectively, and adjust the dose to 0.1 mL four days later. Measure the joint swelling degree on 0, 14 and 21d. Assess the arthritic inde (AI) in different times and record rat weight on 0, 7, 14 and 21d. After 14 d and 21 d, five rats were randomly selected from each group, the ankle joints were prepared for histopathologic study and the level of IL-1β, IL-17 and TNF-α in serum were determined by enzyme-linked immunosorbent assay, respectively. Then the collected data was entered into the database, and SPSS MODELER for data mining analysis was used. RESULTS Three simulation models were established with data mining techniques: the integrated model, C5.0 decision tree model and neural network model, the correct rate for these models were 95%, 90%, 100%, respectively. The cumulative gain curve of these models improved significantly in varying degrees, consistent with AI score and pathological results. CONCLUSION Combined with a practical example (adjuvant arthritis rats model), we describe and elaborate the application process, method, optimization and evaluation of data mining technology in animal modeling. With the introduction of new data processing methods, we can build a serious of forecast evaluation models with different evaluation indicators, and make the right decision by using the information more effectively.