Head and neck cancers are among the most prevalent tumors in the world. Despite advances in the treatment of head and neck tumors, the survival of patients with these cancers has not markedly improved over the past several decades because of our inability to control and our poor understanding of the regional and distant spread of this disease. One of the factors contributing to our poor understanding may be the lack of reliable animal models of head and neck cancer metastasis. The earliest xenograft models in which human tumor cells were grown in immunosuppressed mice involved subcutaneous implantation of human head and neck cancer cell lines. Subcutaneous xenograft models have been popular because they are easy to establish, easy to manage, and lend themselves to ready quantitation of the tumor burden. More recently, orthotopic xenograft models, in which the tumor cells are implanted in the tumor site of origin, have been used with greater frequency in animal studies of head and neck cancers. Orthotopic xenograft models are advantageous for their ability to mimic local tumor growth and recapitulate the pathways of metastasis seen in human head and neck cancers. In addition, recent innovations in cell labeling techniques and small-animal imaging have enabled investigators to monitor the metastatic process and quantitate the growth and spread of orthopically implanted tumors. This review summarizes the progress in the development of murine xenograft models of head and neck cancers. We then discuss the advantages and disadvantages of each type of xenograft model. We also discuss the potential for these models to help elucidate the mechanisms of regional and distant metastasis, which could improve our ability to treat head and neck cancers.