To search for key drug resistance biomarkers and potential chemotherapeutic agents for combination therapy in melanoma patients. The common up-regulated differentially expressed genes were acquired from two cohort studies, GSE91061 and PRJEB23709. Univariate and multivariate Cox regression survival analyses were performed to screen for important prognostic biomarkers. A multi-gene panel including APP, ARHGAP42, GSTA4 and UCHL1, not only plays an important role in prognostic indicators, but also in predicting the probability of resistance to anti-PD-1 in patients with melanoma. Chemotherapy drug response models found five potential drugs, including all-trans retinoic acid, doxorubicin, etoposide, methotrexate and MG-132, were significantly associated with target genes in gene-panel. Molecular docking confirmed that potential drugs have good binding ability to target genes APP, GSTA4 and UCHL1, suggesting that the multi-gene panel is expected to provide assistance for drug resistance prediction and prognosis assessment and combination therapy in patients with anti-PD-1 resistant melanoma.