Background: Impairment of joint proprioception in patients with hypermobility syndrome (HMS) has been well documented. Both joint proprioception and muscle torque are commonly assessed in patients with musculoskeletal complaints. It is unknown, however, if these measures change significantly on repeated application in healthy children and in children with HMS. Aim: To investigate the between-days repeatability of joint proprioception and muscle torque in these groups. Methods: Twenty children (10 healthy and 10 with HMS), aged eight to 15 years, were assessed on two separate occasions (one week apart) for joint kinaesthesia (JK), joint position sense (JPS), and the extensor and knee flexor muscle torque of the knee. JK was measured using threshold to detection of passive movement. JPS was measured using the absolute angular error (AAE; the absolute difference between the target and perceived angles). Knee extensor and flexor muscle torque was normalized to body weight. Results: Intra-class correlation coefficients (ICC) for JK, extensor and flexor muscle torque excellent in both groups (range 0.83 to 0.98). However, ICC values for JPS tests were poor to moderate in the two groups (range 0.18 to 0.56). 95% limits of agreement (LOA) were narrow in both cohorts for JK and muscle torque (indicating low systematic error) but wide for the JPS tests. 95% LOA also demonstrated that the measuring instruments used in this study had low between-days systematic error. Conclusions: Based on ICC and 95% LOA, the repeatability of JK and muscle torque measurements was excellent in both healthy children and those with HMS. The JPS test can only be assessed with poor to moderate repeatability. The use of the JPS test in these children should be undertaken with caution. Copyright © 2008 John Wiley & Sons, Ltd.