Real-time visual feedback might be effective for gait retraining in patients with knee osteoarthritis, to potentially relieve symptoms and postpone knee replacement. In this study, we investigated the effect of various types of real-time visual feedback on a kinetic and a kinematic gait parameter and the different kinematic strategies adopted to reduce knee load. Seventeen healthy subjects walked on an instrumented treadmill while receiving real-time visual kinetic feedback aimed at minimizing the external knee adduction moment (KAdM, reflecting the knee load) or kinematic feedback on the hip internal rotation angle (HIR, a gait modification to reduce the KAdM). Four types of visual feedback (bar, polar plot, color change, graph) were provided. The KAdM decreased by 50 % with kinetic feedback, while kinematic feedback resulted in an HIR increase of 8° but no decrease in KAdM. The degree of change was not influenced by the type of visual feedback. The kinematic changes that reduced KAdM were increased toe-in, increased step width, and decreased hip adduction. Real-time visual feedback can effectively modify gait parameters. Feedback of the KAdM may be more effective in reducing the KAdM than controlling a kinematic parameter that is assumed to unload the knee.