Accumulating epidemiologic and preclinical evidence support the pharmacologic use of a variety of dietary chemicals for the prevention and treatment of cancer. However, it will be challenging to translate these findings into routine clinical practice since phytochemicals have pleiotropic biological activities that have to be balanced for optimal efficacy without unacceptable and potentially unanticipated toxicities. Correctly matching patient populations and settings with optimal, natural product-based phytochemical therapies will require a greater understanding of the specific mechanisms underlying the efficacy, toxicity, and resistance of each agent in a variety of normal, premalignant, and malignant settings. This, in turn, necessitates continued commitment from the basic research community to guide carefully designed and informed clinical trials. The most developed class of anticancer phytochemicals consists of the derivatives of vitamin A called retinoids. Unlike other natural product chemicals currently under study, the retinoids have been extensively tested in humans. Over 30 years of clinical investigation has resulted in several disappointments, but there were some spectacular successes where certain retinoid-based protocols are now FDA-approved standard of care therapies to treat specific malignancies. Furthermore, retinoids are one of the most evaluated pharmacologic agents in the ultra-challenging setting of interventional cancer prevention. This review will summarize the development of retinoids in cancer therapy and prevention with an emphasis on currently proposed mechanisms mediating their efficacy, toxicity, and resistance.