Background Our purpose was to analyze the long-term clinical outcome and to identify prognostic factors after Linac-based stereotactic radiosurgery (SRS) or fractionated stereotactic radiotherapy (FSRT) on patients with brain metastases (BM) from non-small cell lung cancer (NSCLC). Materials and Methods We performed a retrospective analysis of survival on 90 patients who underwent SRS or FSRT of intracranial NSCLC metastases between 04/2004 and 05/2014 that had not undergone prior surgery or whole brain radiotherapy (WBRT) for BM. Follow-up data was analyzed until May 2015. Potential prognostic factors were examined in univariable and multivariable analyses. The Golden Grading System (GGS), the disease-specific graded prognostic assessment (DS-GPA), the RADES II prognostic index as well as the NSCLC-specific index proposed by Rades et al. in 2013 (NSCLC-RADES) were calculated and their predictive values were tested in univariable analysis. Results The median follow-up time of the surviving patients was 14 months. The overall survival (OS) rate was 51 % after 6 months and 29.9 % after 12 months. Statistically significant factors of better OS after univariable analysis were lower International Union Against Cancer (UICC) stage at first diagnosis, histology of adenocarcinoma, prior surgery of the primary tumor and lower total BM volume. After multivariable analysis adenocarcinoma histology remained a significant factor; higher Karnofsky Performance Score (KPS) and the presence of extracranial metastases (ECM) were also significant. The RADES II and the NSCLC-RADES indices were significant predictors of OS. However, the NSCLC-RADES failed to differentiate between intermediate- and low-risk patients. The DS-GPA and GGS were not statistically significant predictors of survival in univariable analysis. Conclusion The ideal prognostic index has not been defined yet. We believe that more specific indices will be developed in the future. Our results indicate that the histologic subtype of NSCLC could add to the prognostic value of specialized future indices. The RADES II index had the highest predictive value in the examined patient cohort.