Chimeric antigen receptors (CAR) can transmit signals akin to those from activated T-cell receptors when bound to a cell surface target. CAR-expressing T cells against CD19 can cause curative effects in leukemia and lymphoma and is approved for clinical use. However, no CAR-T therapy is currently approved for use in solid tumors. We hypothesize that the resistance of solid tumors to CAR-T can be overcome by similar means as those used to reactivate tumor-infiltrating T lymphocytes (TIL), for example, by cytokines or immune checkpoint blockade. Here we demonstrate that CAR-T cells directed against HER2 can kill uveal and cutaneous melanoma cells in vitro and in vivo. Curative effects in vivo were only observed in xenografts grown in a NOD/SCID IL2 receptor gamma (NOG) knockout mouse strain transgenic for human IL2. The effect was target-specific, as CRISPR/Cas9-mediated disruption of HER2 in the melanoma cells abrogated the killing effect of the CAR-T cells. The CAR–T cells were also able to kill melanoma cells from patients resistant to adoptive T-cell transfer (ACT) of autologous TILs. Thus, CAR-T therapy represents an option for patients that do not respond to immunotherapy with ACT of TIL or immune checkpoint blockade. In addition, our data highlight the use of IL2 transgenic NOG mice as models to prove efficacy of CAR-T-cell products, possibly even in a personalized manner. Significance: These findings demonstrate that a novel humanized mouse model can help clinical translation of CAR-T cells against uveal and cutaneous melanoma that do not respond to TIL therapy or immune checkpoint blockade.