Selenium (Se) is known to prevent from several cancers, while iron (Fe) is known to be associated with high risk of cancers. The role of Se on colon carcinogenesis was investigated in an animal model induced by azoxymethane (AOM) and dextran sodium sulfate (DSS) in low Fe mice. Six-week old ICR mice fed on a low Fe diet (4.5 ppm Fe; generally 10 times lower than normal Fe) with three different Se (0.02, 0.1 or 0.5 ppm) levels for 24 weeks. The animals received weekly three ($0{sim}2^{nd}$ weeks) i.p. injections of AOM (10 mg/kg RW), followed by 2% DSS with drinking water for 1 week to induce the colon cancer. There were five experimental groups including vehicle, positive control (normal Fe level, AOM/DSS), Low Fe (LFe) + AOM/DSS+Low Se (LSe), LFe + AOM/DSS + medium Se (MSe) and LFe + AOM/DSS + high Se (HSe) groups. HSe group showed a 66.7% colonic tumor incidence, MSe group showed a 69.2% tumor incidence, and LSe group showed a 80.0% tumor incidence. The tumor incidence was negatively associated with Se levels of diets. Tumor multiplicity in Hse group was significantly low compared to the other groups (p < 0.05). With increasing Se levels of diets, the primary anti-proliferating cell nuclear antigen (PCNA)-positive cells were decreased and apoptotic bodies were increased in a dose-dependent manner. Se-dependent glutathione peroxidase activity and its protein level were dependent on the levels of Se of diets. Malondialdehyde level in liver was lowest in Hse group among experimental groups. These findings indicate that dietary Se is chemopreventive for colon cancer by increasing antioxidant activity and decreasing cell proliferation in Fe-deficient mice.