The present study was conducted to assess the inhibitory effects of EGCG (epigallocatechin-3-gallate) on NMBA-induced rat esophageal tumorigenesis and to seek the potential mechanisms. In experiment I, 81 F344 rats were randomly divided into seven experimental groups according to the different regiments of NMBA 1 mg/kg subcutaneously (s.c.) and EGCG 4 mg/kg or 10 mg/kg orally or intraperitoneally (i.p.). The experiment was terminated at 24 weeks. In experiment II, 48 rats were allocated into two groups, each group contained 24 rats, in which the rats were injected with NMBA 1 mg/kg only or a combination of NMBA 1 mg/kg and EGCG 4 mg/kg i.p. Six rats from each group were sacrificed at the 12th, 16th, 20th and 24th week, respectively. The expression of cyclin D1 and cyclooxygenases (COX-2 and COX-1) was detected using semi-quantitative RT-PCR, and the production of prostaglandin E2 (PGE2) was measured by ELISA. In the groups which were treated with EGCG at a dose of 4 mg/kg i.p., or 10 mg/kg both orally and i.p., the mean number of tumors per rat was significantly reduced to 48, 56 and 61%, respectively (p<0.05). The incidence rate of esophageal carcinomas in the rats that were treated with EGCG 4 mg/kg i.p., was significantly lower than that in the rats which only received NMBA 1 mg/kg (p<0.05). The expression of cyclin D1 and COX-2, and the levels of PGE2 were also decreased by EGCG treatment. These results indicated that EGCG significantly inhibits the NMBA-induced rat esophageal carcinogenesis and it inhibitory effects may partly target cyclin D1 and COX-2 expression, and PGE2 production.