Abstract: There are many unsolved problems that computers could solve if the appropriate software existed. Many real time problems are currently unsolvable such as Flight control systems for aircraft, automated manufacturing systems, and sophisticated avionics systems, not because current computers are too slow or have too little memory, but simply because it is too difficult to determine what the program should do. If a computer could learn to solve the problems through trial and error, that would be of great practical value. Reinforcement Learning is an approach to machine intelligence that combines two disciplines: Dynamic Programming and supervised learning to successfully solve problems that neither discipline can address individually. Encouraged by this emerging technique, this document briefly reviews the basic study of reinforcement learning, its various techniques and applications of it in various fields.