The area of artificial intelligence with its incessant development in all over every sector releases new research domain throughout the decades. AI has entered to the new evolving stage with the combination concept of deep neural networks and reinforcement learning termed as deep reinforcement learning. Though in the last few years, a surge of interest is observed in the field of deep neural networks in the analysis of brain function, deep reinforcement learning has gathered profound research interests reviewing the relationship among learning, representation and decision-making which delivers innovative set of research tools to the brain science with novel proposition. In medical science where patients’ disease diagnosis and treatment greatly depend on huge data analysis, different machine learning algorithms specially reinforcement learning plays a vital role in offering treatment through breakdown of enormous dataset. Introduction: Machine Learning is a science of preparing computers to act like humans by providing lots of data to the machines so that they can learn various tricks by themselves from the data. There are three types of Machine Learning. These are supervised learning, unsupervised learning and reinforcement learning. In supervised learning, machine is provided with huge amount of training data where each input specifies an output. From this training data set, machine builds a model so that it can classify the unknown data based on the data it is trained. In unsupervised learning, model is given as a dataset which is neither labelled nor classified. The machine will identify the patterns from the dataset without human supervision. Reinforcement Learning is an area of machine learning where a software agent needs to take actions on the environment in order to maximize the rewards. Researchers nowadays are trying to launch a new idea named as Deep Reinforcement Learning which is basically a combination of two terms Deep Learning and Reinforcement Learning. After undertaking incessant research analysis, it is proved that Deep Reinforcement Learning achieved success in the game AlphaGo where the machine defeated European Go champion by 5 games to 0. They have proposed a new algorithm AlphaGo where the concept of both Deep Neural Networks and Reinforcement Learning has been used. In this approach, these deep neural networks are trained by a novel combination of supervised learning from human professional games, and reinforcement learning from games of self-play [1]