Abstract: This paper presents a set of test methods for evaluating the performance of active knee-assisted exoskeleton. A motor-driven assist model is established to calculate the theoretical assist efficiency of exoskeleton knee joints.Six healthy examinees are selected to measure comprehensive metabolic indexes of the human body and lower limb muscle activity under the condition of wearing and not wearing exoskeleton.The results of the comprehensive metabolic cost test of the human body showthat, compared with the experiment without wearing the exoskeleton, examinees wearing the exoskeleton see an increase of oxygen consumption of the human body by 3.21% on average, an increase of exhalation of carbon dioxide by 1.27% on average, and an increase of net metabolic power per unit mass by 2.80% on average. The results of the lower limb muscle activity test show that, compared with the experiment without exoskeleton, wearing exoskeleton can effectively reduce the muscle activity of rectus femoris, vastus lateralis and vastus medialis. A controlled weight-bearing test is designed, and the results prove that the weight of the exoskeleton itself cannot be ignored, which increases the comprehensive metabolic cost of the examinees.The exoskeleton can effectively reduce the activity of the wearer’s lower limb muscles, which is basically in line with the theoretical assisting efficiency of exoskeleton knee joints, and the overall power-assisting efficiency is not obvious due to the weight of the wearer.