This study introduces and validates a multi-joint dummy robot designed specifically for the performance evaluation of upper-limb wearable robots. Current evaluation methods predominantly rely on biological signals or metabolic costs, which lack objectivity and reliability. The dummy robot mimics human arm movements and performs repetitive test motions, providing consistent experiment environment, for both cases with and without the wearable robot attached to assess its assistive effectiveness. Experimental results demonstrate the dummy robot's capability generating test motion accurately and repetitively is validated by root mean square error (PRMSE) of 2.922%, and 2.618% for trajectory tracking. Additionally, the muscular strength support performance index (MSSPI) quantifies the workload reduction facilitated by the wearable robot, evaluated at 21.5J, providing a quantitative assessment of the robot's effectiveness. This study proposes a new methodology for effective evaluation of the performance of upper-limb wearable robots, potentially advancing assistive robotic technologies.