This paper investigated the imitation of human motions by a NAO humanoid robot which can be regarded as a human-robot interaction research. In this research, first, human motion is captured by a Kinect 3-dimentional camera through a Robot Operating System (ROS) package. Captured motion is then mapped into the robot’s dimension due to the differences between human and humanoid robot dimensions. After performing the mapping procedure, the solution of both forward and inverse kinematic problem of the robot are solved. To this end, a “Distal” form of forward kinematics solution of the NAO humanoid robot is computed and based on the latter form an analytical inverse kinematics solution for the whole-body imitation purpose is used. The foregoing issue, as one of the contributions of this paper, can be regarded as one of the main reason for obtaining a smooth imitation. In order to keep the robot’s stability during the imitation, an ankle strategy based on a Linear Inverted Pendulum Model (LIPM) and the Ground projection of the Center of Mass (GCoM) criteria is introduced. Moreover, the latter LIPM is controlled by a Proportional-Integral-Derivative (PID) controller for two cases, namely, double and single support phases. Considering the limitation on the motion capture device, from experimental and simulation results obtained by implementing the proposed method on a NAO-H25 Version4 it can be inferred that the robot exhibits an accurate, smooth and fast whole-body motion imitation.