Given the fact that the applications of robot arm have become increasingly extensive, dynamic peg insertion on the production line has also been applied massively. Through robot position control, force measurements feedback, and reinforcement learning, the requirements of operation personnel and the chance of peg damage during insertion can be greatly reduced. The purpose of this thesis is the robot arm peg insertion control application. Combining robotics and reinforcement learning into robot systems, simulation of skew screws adjustment of human wrist can be achieved. In this paper, we design a simple integration interface to visualize the contact forces and torques. Furthermore, using reinforcement learning technique, combined with remote center of motion control, it is possible to train the robot to learn how to adjust the skew of screw like human wrist do.