This study proposes a quantitative evaluation method for assessing active wearable assistive devices that can efficiently support the human body. We utilize a humanoid robot to simulate human users wearing assistive devices owing to the robot’s advantages such as quantitative torque measurement from the sensors and highly repeatable motion. In this study, we propose a scheme for estimating the supportive torques supplied by a device, called stationary torque replacement. To validate the reliability of this evaluation method using a humanoid robot, we also conducted measurements of human muscular activity during assisted motion. The analysis of the measured muscle activity revealed that a humanoid robot closely simulate the actual usages of devices of an assistive devices. Finally, we showed the feasibility of the evaluation method through an experiment with the humanoid robot platform HRP-4 and the Muscle Suit active assistive device. With the method proposed, the supportive effects of an assistive device can be quantitatively measured in terms of the static supportive torque acting directly on the body of a simulated human user.