Many beverages include bioactive components and energy but are frequently not considered in diet quality estimations. We examined the association of a healthy beverage score (HBS) with incident frailty in older adults from the Seniors-ENRICA-1 cohort. We used data from 1900 participants (mean ± SD age 68.7 ± 6.4 years, 51.7% women), recruited in 2008–2010 and followed-up until 2012 assessing food consumption at baseline with a validated diet history. The HBS was higher for increasing consumption of low fat milk, tea/coffee, lower consumption of whole milk, fruit juice, artificially sweetened beverages, sugar-sweetened beverages, and moderate intake of alcohol. Frailty was considered as having ≥3 criteria: exhaustion, low-physical activity, slow gait speed, weakness, and weight loss. We performed logistic regression analyses adjusted for potential confounders. During a 3.5 y mean follow-up, 136 new cases of frailty occurred. Compared to the lowest sex-specific HBS tertile, the fully adjusted odds ratio (95% confidence interval) of frailty was 0.59 (0.38, 0.92) in the intermediate tertile, and 0.52 (0.31, 0.88) in the highest tertile, p trend = 0.007. Results for slow gait speed were 0.79 (0.58, 1.07) and 0.71 (0.51–0.99), p trend = 0.033. Therefore, adherence to HBS was inversely associated with incident frailty and slow gait speed. HBS can help on the beverage quality evaluation, highlighting beverage importance as contributors to diet and to health.