Gait speed is a recommended geriatric assessment of physical performance, but may not be regularly examined in clinical settings. We aimed to investigate whether quadriceps strength tests demonstrate similar predictive ability for incident falls as gait speed in older women. We investigated 135 female volunteers aged mean ± SD 76.7 ± 5.0 years (range 70–92) at high risk of fracture. Participants completed gait speed assessments using the GAITRite Electronic Walkway System, and quadriceps strength assessments using a hand-held dynamometer (HHD). Participants reported incident falls monthly for 3.7 ± 1.2 years. N = 99 (73%) participants fell 355 times during the follow-up period (mean fall rate 83 per 100 person years). We observed a reduced odds ratio for multiple falls (0.83, 95% CI 0.70–0.98) and a reduced hazard ratio for time to first fall (0.90, 95% CI 0.83–0.98), according to quadriceps strength. There was also a significantly shorter time to first fall for those with low quadriceps strength (<7.0 kg; lowest tertile) compared with those with normal quadriceps strength (estimated means [95% CI] 1.54 [1.02, 2.06] vs. 2.23 [1.82, 2.64] years; P = 0.019), but not for those with low (<1.0 m/s) vs. normal gait speed (P = 0.15). Quadriceps strength is a significant predictor of incident falls over three years amongst community-dwelling older women at high risk of fracture. Quadriceps strength tests may be an acceptable alternative to gait speed for geriatric assessments of falls risk.