This study seeks to analyze the effectiveness of inventory management using the Q-Learning algorithm to identify the optimal quantity of spare parts to order for each period, minimizing the costs and downtimes related to stock management and operating smoothly without a hitch. This paper presents a new approach based on Q-Learning, a type of reinforcement learning that allows intelligent decision-making and models the inventory structure as a Markov decision system. This mechanism is to learn an optimal policy of restocking, respecting the stock levels under the (s, S) policy operation replenishment. So, the objective of this article is to simulate the results of the Q-Learning algorithm approach that demonstrates efficiency in reducing costs and improving service levels by adapting to demanding dynamic changes, epitomizing the revolution and the new advancements technologies in inventory management by Artificial Intelligence that enable and provide practical and efficient insights into the implementation of Q-Learning industry settings.