Objective To construct a rheumatoid arthritis (RA) model of female SD rats with complete Freund's adjuvant (CFA), and to investigate the therapeutic effects of celecoxib on RA with different administration methods. Methods Twenty-four female SD rats were randomly divided into normal control group and low-, medium-, high-dose model groups (n=6), and 0.05, 0.1 and 0.2 mL of CFA (10 mg/mL) were injected into the right hind limb plantar region of the rats to induce RA models, and the appropriate dose was screened. Another 24 female SD rats were randomly divided into normal control group and 3 experimental groups after the establishment of the optimal inflammation model (model control group, celecoxib oral group, and celecoxib injection group) (n=6). The celecoxib oral group was given commercial celecoxib capsules, 20 mg each time, once a day by gavage; the celecoxib injection group was given a celecoxib suspension, administered once a week at a dose of 20 mg/kg. The foot volume of rats was measured in each group, and the clinical scores were performed. The levels of 3 inflammatory factors (interleukin[IL]-1β, IL-6, and tumor necrosis factor α[TNF-α]) in serum of rats in each group were detected by enzyme-linked immunosorbent assay, and the ankle joint tissues of rats were stained with hematoxylin-eosin staining to observe the pathological changes. Results In female SD rats, moderate inflammatory reaction of RA could be better simulated by injecting 0.1 mL CFA (10 mg/mL) into the plantar part of the foot. After CFA injection, the feet gradually become red and swollen, and the systemic clinical score and foot arthritis index score were both increased. The peak of inflammation appeared on day 19 after immunization. The foot volume, systemic clinical score, foot arthritis index score, the concentrations of serum IL-1β, IL-6 and TNF-α of the rats in the celecoxib oral group and injection group were significantly lower than those in the model control group (all P<0.05), and the pathological changes of ankle joint were all alleviated. The systemic clinical score and foot arthritis index score of the rats in the celecoxib injection group were significantly lower than those in the oral group (both P<0.05), and the pathological changes of ankle joint were mild in the injection group. Conclusion Injection of 0.1 mL CFA (10 mg/mL) into the plantar region of the foot can cause moderate inflammation reaction in female SD rats, and it can better simulate the pathological process of RA. Injecting celecoxib suspension once a week has a better effect than taking celecoxib orally once a day on RA in model rats.