Background: The updated Graded Prognostic Assessment for Lung Cancer Using Molecular Markers (Lung-molGPA) index will provide more accurate prediction of expected survival for patients with advanced non-small cell lung cancer (NSCLC) with brain metastases (BM). This retrospective study aims to evaluate optimal timing of cranial radiotherapy (RT) in NSCLC patients with BM stratified by Lung-molGPA index. Methods: This study retrospectively screened patients with NSCLC and initially diagnosed BM in our cancer center between February 2012 and July 2018. These enrolled patients were scored and stratified based on lung-molGPA index. The primary endpoint was overall survival time (OS), defined as the interval from the initiation of BM diagnosis to death caused by any reason or the last follow-up. And the secondary endpoint was progression free survival (PFS), referred to the interval from the initiation of BM diagnosis to disease progression or death caused by any reason. The effect of radiotherapy timing on survival time in groups stratified by lung-molGPA index was evaluated. The early RT was defined as receiving RT within three months after initial diagnosis of BM with no progression of BM prior to RT. And the delayed RT was defined as receiving RT after progression of BM or initial diagnosis of BM three months later. Results: Overall, 471 patients were enrolled in our study, including 260 patients receiving RT. The median follow-up time was 48 months. In entire cohort, median OS time and PFS were 15.0 and 9.0 months, respectively. In lung-molGPA0-2 group, OS of patients with or without RT was no statistical difference (HR, 0.78, 95%CI, 0.59-1.04, P= .071). While in lung-molGPA2.5-4 group, the addition of RT could prolong OS and PFS of patients (OS: HR, 0.51, 95%CI, 0.39-0.67, P< .001; and PFS: HR, 0.66, 95%CI, 0.51-0.85, P< .001). Moreover, in patients scored as 2.5-4.0, the median OS of early RT and delayed RT was 24.0 months and 15.0 months, respectively, and OS of patients receiving early RT was significantly superior to patients receiving delayed RT (HR, 0.50, 95%CI, 0.31-0.79, P< .001). While for patients scored as 0-2, no statistical difference was between OS of patients receiving early RT and delayed RT (HR, 0.75, 95%CI, 0.51-1.10, P= .115). Conclusions: With regards as patients scored 2.5-4.0 based on lung-molGPA index, early addition of cranial radiotherapy could benefit patients’ survival. While for patients scored 0-2, the survival benefit from cranial radiotherapy was limited.