Although numerous studies have assessed the effect of foods and nutrients on colorectal carcinogenesis, few studies have investigated human eating behavior in relation to risk of colorectal cancer. In our study, we assessed whether the reported behavior of eating anything at anytime influenced colorectal cancer risk and related plasma biomarkers. We prospectively followed up 55,540 women in the Nurses' Health Study who were aged 48–73 years, had no history of cancer, ulcerative colitis or diabetes and responded to the item “I eat anything I want, anytime I want” in the 1994 questionnaire. We also analyzed blood samples for 1,994 women, which were collected in 1989–1990. During 12 years of follow-up, 552 colorectal cancer cases were documented. After adjusting for age, smoking, body mass index, physical activity, red and processed meat and other known risk factors for colorectal cancer, women who reported eating anything at anytime experienced an increased risk of colorectal cancer (relative risk = 1.28, 95% confidence interval = 1.06–1.56) compared to those who did not report this behavior. In addition, reporting eating anything at anytime was associated with higher fasting plasma levels of insulin (p= 0.04) and C-peptide (p= 0.05). In conclusion, reports of eating anything at anytime are associated with an increased risk of colorectal cancer in this large prospective cohort study, independent of other potential risk factors for colorectal cancer.