The CHADS2, CHA2DS2-VASc, and R2CHADS2 scores are well-known predictors of stroke caused by atrial fibrillation (AF), but no studies have evaluated their use for stratifying all-cause mortality risk in patients discharged for systolic heart failure (SHF) with or without AF.This study analyzed data in the Taiwan Society of Cardiology-heart failure with reduced ejection fraction (TSOC-HFrEF) registry. These data were obtained by a prospective, multicenter, observational survey of patients treated at 21 medical centers in Taiwan after hospitalization for acute, pre-existing or new onset SHF from May, 2013 to October, 2014. During 1 year follow-up, 198 patients were lost follow-up, and final 1311 (86.8%) patients were included for further analysis. During the follow-up period, 250 (19%) patients died. Multivariate analysis revealed that body mass index, thyroid disorder, valvular surgery history, chronic kidney disease (CKD), and scores for CHADS2, CHA2DS2-VASc, and R2CHADS2 were significant independent predictors of mortality in the overall population of SHF patients (all P < .05) The c-indexes showed that CHADS2, CHA2DS2-VASc, and R2CHADS2 scores were significantly associated with mortality in SHF patients with or without AF (all P < 005). However, R2CHADS2 had significantly higher accuracy in predicting mortality in all SHF patients compared with CHADS2 and CHA2DS2-VASc (DeLong test, P < .0001), especially in SHF without AF (DeLong test, P = .0003).Scores for CHADS2, CHA2DS2-VASc, and R2CHADS2 can be used to predict 1-year all-cause mortality in SHF patients with or without AF. For predicting all-cause mortality in SHF patients, R2CHADS2 is more accurate than CHADS2 and CHA2DS2-VASc.