This study aimed to identify clinical factors associated with survival and to develop a prognostic tool in patients receiving radiosurgery alone for very few cerebral metastases from melanoma. Ten characteristics of 69 patients treated with radiosurgery alone for 1-3 cerebral metastases from melanoma were retrospectively analyzed for survival. Serum lactate dehydrogenase levels before radiosurgery, Karnofsky performance score, maximum diameter of all irradiated cerebral lesions, and extracranial lesions were significantly associated with survival and included in the tool. Twelve-month survival rate (in %) divided by 10 was calculated for each of these four factors. The four scores were summed resulting in total scores ranging 9-22 points. Based on the 12-month survival rates, three groups were formed: less than 15 points (group I, n = 25); 15-20 points (group II, n = 34); and more than 20 points (group III, n = 10). Corresponding 12-month survival rates were 10 ± 6%, 51 ± 9% and 90 ± 9%, respectively (P < 0.001). In group I, death within 12 months following radiosurgery occurred mostly due to extracranial progression, whereas cerebral progression was the major cause of death in group II. In group III, only one of 10 patients died within 1 year. This new prognostic tool helps predict the survival time following radiosurgery of very few cerebral metastases from melanoma. An individual treatment approach should consider a patient's survival time and the most likely cause of death (cerebral or extracranial progression).