The proposed reinforcement learning (RL) based framework leverages the dynamic nature of electric vehicle (EV) charging demands and the adaptability of RL algorithms to make efficient decisions on charging resource allocation. Based on the present condition of the charging station and the anticipated future demand, the RL agent learns to make the best charging decisions by rephrasing the resource allocation problem as a Markov decision process (MDP). The study looks at and contrasts a number of RL algorithms, including deep Q-network (DQN) and Q-learning and actor-critic methods, to identify the most suitable algorithm for the resource allocation task in electric vehicle charging stations (EVCSs). It also considers the integration of real-time data, such as EV arrival patterns, charging profiles, and grid conditions, to enhance the RL agent's decision-making capabilities. To validate the proposed approach, extensive simulations are conducted using realistic charging scenarios and performance metrics, including charging station utilization, waiting time, and energy efficiency. The results demonstrate the effectiveness of the RL-based resource allocation approach, which outperforms traditional static allocation methods in terms of resource utilization and overall charging system performance. This paper contributes to the growing field of EV charging infrastructure optimization by providing a comprehensive study on real-time resource allocation using RL. The findings pave the way for the implementation of intelligent and adaptive charging systems, facilitating the seamless integration of EVs into the existing power grid and fostering sustainable transportation.