Abstarct: With the advancement of technology, the arrival of robots to various aspects of human life has become inevitable. In recent years, the effort of the robotics community has focused on the development of nature-inspired robots, in particular humanoid biped robots. Since living organisms, especially humans, have unique characteristics such as precision, speed, and sustainability in movement; therefore, the creation of these features in robots is increasing day by day. On the other hand, if the movement of robot looks like a human being, it will create a sense of security and comfort, and will increase the productivity of the person and the quality of the service. Man-made robots made to this day, despite the extreme stability, have a significant difference in the pattern of gait. Therefore, in this study, we decided to consider the idea of direct use of human walking data in the motion of biped robots. The data recorded in this study was carried out by motion capture method with the aid of high frequency cameras. After processing data and converting those to the state space, a Lagrange model of the 5-lixnk robot were developed, after calculating the inverse dynamics, the torques needed to produce the similar motion to human were reported. Sliding mode and a feedback linearization controller were designed to produce this movement and the desired trajectory was successfully tracked. After that, human model moments were obtained by Opensim software and compared with the robot model. The large difference in torque, despite the uniformity of movement (the trend was almost the same), was associated with the idea that one of the reasons for not having the optimized torque values was the absence of ankle joint. Thus, a 7-lixnk model with ankles and control was introduced and it was observed that the torque is much similar to the actual human torque.