There has been a long fascination with immunizing against cancer in general and against melanoma in particular. The earliest melanoma vaccines were formulated from autologous or allogeneic melanoma cells. Subsequently, molecularly defined vaccines made from proteins, peptides, or gangliosides were developed and, recently, DNA-based vaccines are currently being tested. Randomized trials using allogeneic melanoma cells/lysates or gangliosides have not demonstrated a clinical benefit, although in some trials, clinical benefit was seen in large subsets. On the other hand, some clinical trials indicated that vaccine therapy could be associated with a poorer survival. There are several mechanisms now identified by which melanoma cells can avoid detection by T cells, such as loss of expression of antigen or human leukocyte antigen (HLA) molecules, and tumor vascular adhesion molecules. Also, mediators of immune tolerance are now well-characterized such as regulatory T cells and inhibitory cytokines secreted by melanoma cells. Future approaches for immunizing against melanoma will have to incorporate strategies to overcome these barriers. Strategies being tested include depletion of regulatory T cells, immunization in the setting of lymphopenia, and blockade of T-cell inhibitory molecules such as CTLA-4. Once a relevant clinical response can be induced, it should be possible to develop validated immunoassays that will direct future melanoma vaccine development.