The number of mobile robots and their use inthe world is increasing. Robotic robots can move on the ground indoors or outdoors while doing a set of tasks. Today, mobile robots are used in factories, hospitals and for office automation. Although the intelligence and robot capabilities of the robot are improving, however, with the expansion of functional areas, it is becoming a new trend for multiple robots to work together to perform complex tasks.In this thesis, a new approach based on the step backward control method and the leader-follower method for arranging two-wheel robots has been presented. Also, in this method, two states of existence and absence of noise have been considered. Implementation of the proposed method on a sample system has been done and the results of simulation have been investigated and analyzed. Investigating the real-time paths of robots using the formation controller shows that the follower robot can follow the leader well and maintain the desired distance and angle between himself and the leader. The robot error of the follower converges over time to zero. The real time investigation of linear velocity and angular velocity using the proposed controller and regardless of noise indicates that the angular and linear velocity of the leader can be followed by followers. With time, the angular velocity of each follower can quickly aspire to the leader's angles. Also, it can be seen that with time, the linear velocity of each follower can be converge to the leader's linear velocity. In addition, using the proposed approach, tracking errors of followers are approaching zero over time. Investigations in the non-use of the kalman filter also indicate that the effect of noise on the path of the robots is very high and the follower, can not follow the desired path. Keywords: double wheel robot, leader-follower, noise, formation, backstepping control, kalman filter.