This study is aimed at developing an intelligent factory control system for improving environment quality and reducing electricity consumption. By automating intelligent equipment and leveraging internet networks, the system enables the remote monitoring and management of environmental conditions. We combine the fuzzy method and deep reinforcement learning (DRL) to handle complex factory data and optimize decision-making. The fuzzy method uses fuzzy sets and rules to generate accurate outputs from the data. On the other hand, the DRL system learns optimal policies by interacting with the environment using environment quality, central air conditioner (AC), and alarm data. Hardware implementation uses an ESP32-S microcontroller to send data to Google’s Firebase cloud for seamless management and monitoring through a mobile app or website. The study involves developing 36 fuzzy rules and creating 10 models with different combinations of hidden layers, epochs, and learning rate values. Among the fuzzy inference system (FIS)-DRL modeling results, the fourth model stands out as the preferred option to proceed with the experiment, as it achieves the highest accuracy of 91.16%. Note that this model also exhibits a loss value of 1.64% and an incredibly short inference time of only 3 ms. The proposed system offers benefits such as enhanced energy efficiency and reduced costs, making it ideal for intelligent factories. By optimizing resource usage, it will contribute to sustainable development in various industries.