Lower limb rehabilitation robot has showed great potential in improving human walking ability. The purpose of this study is to compare the differences in human muscle activity of the lower limb while walking with different patterns. In order to meet the individual needs of different kinds of patients with lower limb dysfunction, an end-effector hybrid rehabilitation robot concept with multi-degree freedom is designed based on human factor. Human-robot coupling model is established in the AnyBody Modeling System, inverse dynamic analysis of movement and muscle activity of the human body is simulated. Effects of gait trajectory and the degree of freedom simplification of rehabilitation robot on human physiological parameters are discussed. The results indicate that foot trajectory, ankle posture, and mobility of robot is of vital importance and can be realized according to specific human factor in rehabilitation robot design and trajectory planning. Analysis methods and results are meaningful for lower limb rehabilitation robot design and rehabilitation evaluation.