It is becoming clearly evident that single gene or single environmental factor cannot explain susceptibility to diseases with complex etiology such as head and neck cancer. In this study, we applied the multifactor dimensionality reduction method to explore potential gene-environment and gene-gene interactions that may contribute to predisposition to head and neck cancer in the North Indian population. We genotyped 203 patients with head and neck cancer and 201 healthy controls for 13 functional polymorphisms in genes coding for tobacco metabolizing enzymes; CYP1A1, CYP2A13, GSTM1, and UGT1A7 using polymerase chain reaction-restriction fragment length polymorphism method, real-time polymerase chain reaction quantitative assay, and denaturing high-performance liquid chromatography followed by direct sequencing. We found that GSTM1 copy number variations were the most influential factor for head and neck cancer. We also observed significant gene-gene interactions among GSTM1 copy number variants, CYP1A1 T3801C and UGT1A7 T622C variants among smokers. Multifactor dimensionality reduction approach showed that the three-factor model, including smoking status, CYP1A1 T3801C, and GSTM1 copy number variants, conferred more than fourfold increased risk of head and neck cancer (odds ratio 4.89; 95% confidence interval: 3.15-7.32, p < 0.01). These results support the hypothesis that genetic variants in tobacco metabolizing genes may contribute to head and neck cancer risk through gene-gene and gene-environmental interactions.