OBJECTIVE To study knee muscle activity and static postural sway in girls with generalized joint hypermobility (GJH). METHODS Sixteen girls with GJH and 11 girls with non-GJH (NGJH) aged 14 years, randomly recruited among schoolchildren, participated in this study. GJH inclusion criteria were: Beighton score minimum 6/9 and one hypermobile knee; for NGJH: Beighton score maximum 5/9 and no knees with hypermobility. The participants performed a static two-legged balance test with eyes open (2EO) and eyes closed (2EC) and a one-legged stance test with eyes open (1EO). Postural sway (centre of pressure path length, COPL) was calculated, along with rambling and trembling components. Surface electromyography (sEMG) from the quadriceps (Q), hamstrings (H), and gastrocnemius (G) muscles was recorded, expressed as a percentage of the maximum voluntary EMG (%MVE), and the co-contraction index (CCI) of Q, H, and G muscle activity was calculated. Knee function was self-reported using the Knee Injury and Osteoarthritis Outcome Score for children (KOOS-Child). RESULTS GJH had a significantly lower lateral HQ CCI and a higher medial/lateral HQ CCI ratio in all balance tasks. Group mean EMG varied from 1.3%MVE in Q (during 2EO) to 15.7%MVE in G (during 1EO). GJH had larger postural sway length than NGJH during 2EC (COPL 1.64 vs. 1.37 m/min, p < 0.001). Rambling and trembling components did not differ between groups. CONCLUSIONS Girls with GJH and at least one hypermobile knee performed, compared with NGJH, static balance tasks with higher medial knee muscle activity relative to the lateral activity, and larger postural sway when vision was eliminated. The short- and long-term consequences should be studied further.