Background The aim of this prospective study was to evaluate the association between heart rate variability (HRV) and overall survival of lung cancer patients with brain metastases (LCBM). Methods Fifty-six LCBM patients were enrolled in this study. Five-minute electrocardiograms were collected before the time to first brain radiotherapy. HRV was analyzed quantitatively by using the time domain parameters, i.e., the standard deviation of all normal-normal intervals (SDNN) and the root mean square of successive differences (RMSSD). Survival time for LCBM patients was defined as from the date of HRV testing to the date of death or the last follow-up. Results In the univariate analysis, SDNN ≤ 13 ms (P = 0.003) and RMSSD ≤ 4.8 ms (P = 0.014) significantly predicted poor survival. Multivariate analysis confirmed that RMSSD ≤ 4.8 ms (P = 0.013, hazard ratio = 3.457, 95% confidence interval = 1.303–9.171) was also an independent negative prognostic factor after adjusting for mean heart rate, Karnofsky performance status, and number of brain metastases in LCBM patients. Conclusion Decreased RMSSD is independently associated with shorter survival time in LCBM patients. HRV might be a novel predictive biomarker for LCBM prognosis.