BACKGROUND: Establishing the rat model of ulcerative colitis by 2,4,6-trinitrobenzene sulfonic acid (TNBS)/ethanol method is easy to operate, has long duration and the histological changes are similar to human ulcerative colitis, which is commonly used animal model. However, the optimal dose and modeling times still remain controversial. OBJECTIVE: To establish the rat model of the severe ulcerative colitis by different times of different doses of TNBS/ethanol, and to explore the optimal modeling times. METHODS: Eighty Sprague-Dawley rats, SPF level, were provided by Hunan Slack Jingda Experimental Animal Co., Ltd., and the study was approved by the Experimental Animal Ethics Committee of the First Affiliated Hospital of Guangxi University of Chinese Medicine, approval number: 2013(KF)-E-003. The rats were randomly divided into normal control group, and TNBS/ethanol (100 mg/kg once, 100 mg/kg twice, and 100 mg/kg thrice) group. The rats were intragastrically injected with TNBS/ethanol through rectum for 1, 2, and 3 days continuously. Then ten rats were killed at 3 and 7 days after administration, so as to observe the physiological status, colon morphology, disease activity index and histopathological changes. RESULTS AND CONCLUSION: (1) At 7 days after modeling, the body mass of rats progressively decreased and symptoms gradually exacerbated with the modeling times increased. The disease activity index was on a rise. (2) The body mass in the TNBS/ethanol thrice group was significantly lower than that in the other groups (P < 0.05), and the disease activity index and gross scores were significantly higher than those in the other groups (P < 0.05). (3) The pathological results showed that the TNBS/ethanol thrice group caused the lesions located in the colonic mucosa and muscularis, the goblet cells lost, irregular crypts in shape, disordered arrangement and the most of visible ulcers, which were more consistent with the pathological manifestations of severe ulcerative colitis than the other two groups. (4) In summary, in the TNBS/ethanol modeling, using 100 mg/kg TNBS/ethanol thrice is more consistent with the model of severe ulcerative colitis. [ABSTRACT FROM AUTHOR]