Background There is a lack of agreement about applicable instrument to screen frailty in clinical settings. Aims To analyze the association between frailty and mortality in Finnish community-dwelling older people. Methods This was a prospective study with 10- and 18-year follow-ups. Frailty was assessed using FRAIL scale (FS) (n = 1152), Rockwood’s frailty index (FI) (n = 1126), and PRISMA-7 (n = 1124). To analyze the association between frailty and mortality, Cox regression model was used. Results Prevalence of frailty varied from 2 to 24% based on the index used. In unadjusted models, frailty was associated with higher mortality according to FS (hazard ratio 7.96 [95% confidence interval 5.10–12.41] in 10-year follow-up, and 6.32 [4.17–9.57] in 18-year follow-up) and FI (5.97 [4.13–8.64], and 3.95 [3.16–4.94], respectively) in both follow-ups. Also being pre-frail was associated with higher mortality according to both indexes in both follow-ups (FS 2.19 [1.78–2.69], and 1.69 [1.46–1.96]; FI 1.81[1.25–2.62], and 1.31 [1.07–1.61], respectively). Associations persisted even after adjustments. Also according to PRISMA-7, a binary index (robust or frail), frailty was associated with higher mortality in 10- (4.41 [3.55–5.34]) and 18-year follow-ups (3.78 [3.19–4.49]). Discussion Frailty was associated with higher mortality risk according to all three frailty screening instrument used. Simple and fast frailty indexes, FS and PRISMA-7, seemed to be comparable with a multidimensional time-consuming FI in predicting mortality among community-dwelling Finnish older people. Conclusions FS and PRISMA-7 are applicable frailty screening instruments in clinical setting among community-dwelling Finnish older people.