Abstract To examine whether self-selected hydrogen-rich water consumption was associated with 6-month changes in physical function among community-dwelling older adults attending community salons. This prospective observational study included 128 older adults in Hiroshima, Japan. Participants were classified as consumers or non-consumers at baseline and completed repeated assessments of the 30-second chair stand test, Timed Up and Go, usual gait speed, one-leg stance time, and grip strength. Primary analyses used analysis of covariance with robust standard errors, adjusting for baseline value, age, sex, fall history, and walking-aid use; mixed-effects and overlap-weighted models were sensitivity analyses. Compared with non-consumers, consumers had better adjusted 6-month chair-stand performance (1.85 repetitions, 95% confidence interval 1.11 to 2.59), gait speed (0.07 m/s, 0.04 to 0.10), one-leg stance time (1.78 s, 0.43 to 3.13), and grip strength (0.80 kg, 0.21 to 1.40). Timed Up and Go did not clearly differ (-0.39 s, -1.23 to 0.45). Self-selected hydrogen-rich water consumption was associated with better chair-stand performance and gait speed in this pragmatic community setting, but residual confounding remains possible.