Radiosurgery is conventionally prescribed for brain metastases with a single dose of radiation. Fractionation has been advocated to improve tumour control. A multivariate analysis of prognostic factors including fractionation has been performed in two consecutive prospective radiosurgery protocols with and without fractionation in order to identify an association, if any, between fractionation and survival. A surgically applied stereotactic head frame was used. Radiosurgery planning was based on a contrast-enhanced CT. Sixty-nine patients underwent the two-fraction regimen and 35 patients had a single treatment. Multivariate analysis showed that the presence of extracranial malignancy, performance status, multiple brain metastases, patient gender and the time from the initial treatment to radiosurgery were independent determinants for survival. Fractionation was also an independent determinant with two-fraction patients surviving a median of 30 weeks versus single fraction patients who survived a median of 16 weeks. Fractionated radiosurgery was associated with improved survival and deserves further investigation.