Reinforcement learning theory, which started from a theory of psychology, began to be used as a popular unsupervised learning algorithm. With exploiting deep learning it has overcome the existing limitations and lead the innovation in the field of control, robotics, and artificial intelligence. In this course, based on the basic theory of reinforcement learning, we develop basic skills that can utilize reinforcement learning by defining problems in various fields and coding them directly. To this end, students will learn the traditional core foundation theory of Markov decision process, dynamic programming, temporal difference learning, and algorithms such as Double Q network (DQN), Deep Deterministic Policy Gradient (DDPG), and Trust Region Policy Optimization (TRPO) combined with deep learning. Atari game and Mujoco simulating physics will be used for testing algorithms to have hands-on skills in implementing reinforcement learning with python.