Telomere length is associated with aging-related pathologies. Although the association between telomere length and frailty has been studied previously, only a few studies assessing longitudinal changes in telomere length and frailty exist. Longitudinal cohort study. A subpopulation of the Helsinki Birth Cohort Study consisting of 1078 older adults aged 67 to 79 years born in Helsinki, Finland, between 1934 and 1944. Relative leukocyte telomere length (LTL) was measured using quantitative real-time polymerase chain reaction at the average ages of 61 and 71 years, and at the latter the participants were assessed for frailty according to Fried criteria. The mean ± SD relative LTLs were 1.40 ± 0.29 (average age 61 years) and 0.86 ± 0.30 (average age 71 years) for the cohort. A trend of shorter mean relative LTL across frailty groups was observed at 61 years (P = .016) and at 71 years (P = .057). Relative LTL at age 61 years was significantly associated with frailty: per 1-unit increase in relative LTL, the corresponding relative risk ratio (RRR) of frailty was 0.28 (95% confidence interval [CI] 0.08-0.97), adjusting for several confounders. Also, LTL at age 71 years was associated with frailty (RRR 0.18, 95% CI 0.04-0.81) after adjustment for sex, age, and adult socioeconomic status, but further adjustment attenuated the association. No associations between telomere shortening and frailty were observed during the 10-year follow-up. Shorter relative LTL was associated with frailty in cross-sectional and longitudinal analyses, but telomere shortening was not, suggesting that short LTL may be a biomarker of frailty.