Recommendation systems are highly employed information filtering systems designed to suggest items to the user based on the user-s interest in the items. Classic recommendation systems suggest the items only depend on the informational relationship between the user and items. Such systems lack the awareness of contextual information of the user. The significance of the contextual information focuses on the associated information of the user’s situation while interacting with the item. Context-aware recommender systems produce more pertinent recommendations by adapting the system to the specific contextual situation of the user. Existing learning models for contextual recommendation are unable to work on dynamic environments and continuously online learning systems. Reinforcement learning owns the ability to learn long-term rewards, balance exploration, and exploitation, and continuously learn online, whereas deep learning models automatically extract complex data representations from the high-dimensional database. Introducing deep reinforcement learning in contextual recommendations yields real-time recommendations by adapting to the real-world dynamic environment and exhibits superior performance over other learning models.