Abstract: An overall motion control strategy of the exoskeleton is proposed, and the motion planning and PD (proportional-derivative) control of the exoskeleton's swing leg are introduced especially. As the basis of the motion planning of exoskeleton's swing leg, the detection of human-machine posture error at ankle joint can be achieved by banding the exoskeleton and the wearer with each other at waist and ankle, thus the limitation on wearer's movement and the complexity of control system can be reduced. Correctness of the exoskeleton dynamical model which is the basis for designing the swing leg PD controller is verified with Adams and Matlab. Kalman filter adopted to predict the movement of the wearer's swing leg can improve the control accuracy under the limited sampling frequency. The experiment results indicate that this method can efficiently identify the movement intention of wearer's swing leg, and drive the exoskeleton's hip and knee joints to precisely track the wearer's ankle joint trajectory. The method can be used to control the exoskeleton's swing leg.