OBJECTIVE: To investigate the correlation of the polymorphisms of EC-SOD and GSTM1, as well as smoking, with oral cancer risk. METHODS: A case-control study and polymerase chain reaction (PCR) technique was used to identify the genotype of EC-SOD and GSTM1 in 600 oral cancer cases and 600 controls. Logistic regression was conducted to evaluate the odds ratio of relative gene and the interaction between the genes and smoking with oral cancer. RESULTS: The frequencies of EC-SOD genotype in cases and controls were 61. 17% and 78.50% for C/C, and 38.33% and 21.50% for C/G respectively (P < 0.01). EC-SOD (C/G) correlated with susceptible to oral cancer (OR = 2.27, 95% CI 1.73 - 4.02). The frequency distribution of GSTM1 ( - ) was significantly different in patients (69.17%) and controls (44.17%) (P < 0.01). GSTM1 ( -) increased the risk of developing oral cancer (OR = 2.84, 95% CI 1.95 - 4.47). Combined polymorphism analysis showed that the percentage of EC-SOD (C/G)/GSTM1 (-) for oral cancer and control groups was 30.50% and 6.67% respectively) (P < 0.01). The risk of developing oral cancer in those with EC-SOD(C/G)/GSTM1 (-) genotype was significantly higher than that with EC-SOD (C/C)/GSTM1 (+) genotype (OR = 8.16, 95% CI 3.73 - 12.91). There were significant differences in smoking rate between the two groups (P < 0.01), smoking correlated with susceptible to oral cancer (OR = 2.66, 95% CI 1.45-4.36), and statistical analysis suggested an interaction between smoking and EC-SOD(C/G)/GSTM1(-) genotype, which increase the risk of oral cancer (OR = 25.11, 95% CI 12.37 - 36.62). CONCLUSION: EC-SOD (C/G) and GSTM1 (-) are risk factors in oral cancer. Smoking is also related to the susceptibility to oral cancer. There may be a synergetic interaction among EC-SOD (C/G), GSTM1 (-) and smoking on the elevated susceptibility of oral cancer.