Abstract Background Muscle strength and balance are major modifiable factors of falls in older adults, but their associations with falls in middle-aged adults are underinvestigated. We aimed to examine the association of baseline and change in leg muscle strength (LMS) and balance with the incidence of falls in a cohort of middle-aged women. Methods This was a 5-year follow-up of a population-based sample of 273 women aged 36–57 years at baseline (2011–2012). Data on LMS (by dynamometer) and balance (timed up and go test, step test, functional reach test, and lateral reach test) were obtained at baseline and 5 years later (2017–2018). After 5 years, falls were recorded monthly for 1 year by questionnaire (2017–2019). Negative binomial/Poisson and log-binomial regressions were used as appropriate to assess associations of baseline and change in LMS and balance with any falls, injurious falls, and multiple falls. Results Over 1 year, 115 participants (42%) reported at least one fall. Neither baseline nor 5-year change in LMS and balance measures was associated with the risk of any falls, injurious falls, or multiple falls 5 years later, with or without adjusting for confounders at baseline (incidence rate ratio/relative risk ranging from 0.85 to 1.19, 0.90 to 1.20, and 0.82 to 1.36, respectively; p > .05 for all). Conclusions Baseline or change in LMS and balance measures are not associated with incident falls among middle-aged women. The contributions of environmental and other intrinsic factors such as chronic conditions and gait/mobility problems need to be investigated.