Now effective adjuvant therapy has arrived in melanoma, accurate staging and patient selection to optimize its risk/benefit ratio is crucial. The American Joint Committee on Cancer staging system is the most widely used and validated melanoma staging system, which recently released its 8th edition. We aimed to externally validate the prognostic and discriminatory ability for survival of the 8th edition compared to the 7th edition and evaluate prognostic factors. Prospective database of stage III melanoma (2000-2016). Prognostic factors for melanoma-specific survival and distant metastasis-free survival were analyzed. Survival differentiation of the 7th and 8th edition was assessed with log-rank tests and Cox proportional hazards models. Discriminatory ability was compared using the receiver operating characteristic and Akaike's Information Criterion. Six hundred forty patients were included (median follow-up 59 months). Median melanoma-specific survival was 138 months, distant metastasis-free survival 96 months. Age, Breslow thickness, ulceration of the primary tumor and number of positive lymph nodes (N) were independent prognostic parameters for distant metastasis-free survival and melanoma-specific survival. The 8th edition performed slightly better than the 7th edition in terms of survival discrimination but showed slightly worse distant metastasis-free survival and melanoma-specific survival differentiation between stage IIIA and IIIB. Sentinel node (SN) metastasis size cutoff of 1 mm differentiated survival in both 7th and 8th edition stage IIIA, showing excellent distant metastasis-free survival and melanoma-specific survival for patients with a SN metastasis size <1 mm. The 8th edition performed at least comparably, if not better than the 7th in terms of survival discrimination. However, survival in both 7th and 8th edition stage IIIA melanoma remains heterogeneous. EORTC SN tumor burden criteria can further stratify survival and help patient selection for adjuvant therapy.