Purpose: Hand-held dynamometry (HHD) can be used to evaluate strength when gold-standard isokinetic dynamometry (IKD) is not feasible. HHD is useful for measuring lower limb strength in a healthy population; however, its reliability and validity in individuals with knee osteoarthritis (OA) has received little attention. In this research, we examined the test–retest reliability and validity of HHD in older women with knee OA. We also examined the associations between reliability and symptom and disease severity. Method: A total of 28 older women with knee OA completed knee extension and flexion exertions measured using HHD and IKD. Intra-class correlation coefficients (ICC2,3), standard error of measurement, and minimal detectable change were calculated. Correlation coefficients and regressions evaluated the relationships between inter-trial differences and symptom and disease severity. Results: High test–retest reliability was demonstrated for both exertions with each device (ICC2,3 = 0.83–0.96). Variance between trials was not correlated with OA symptoms. Criterion validity was good (ICC2,3 = 0.76), but extension yielded lower agreement than flexion. Regression analysis demonstrated that true strength can be predicted from HHD measurements. Conclusions: HHD is a reliable tool for capturing knee extension and flexion in individuals with OA. Because of lower agreement, HHD might be best suited for evaluating within-subject strength changes rather than true strength scores. However, gold-standard extension strength magnitudes may reasonably be predicted from regression equations ( r2 = 0.82).