Abstract: Industrial Control Systems (ICS) play a crucial role in modern industrial production, responsible for monitoring and controlling various industrial processes to ensure efficiency, safety, and stability in production. With the development of Industry 4.0 and smart manufacturing, traditional industrial control methods are increasingly inadequate to handle the growing complexity and dynamic nature of production environments. Deep Reinforcement Learning (DRL), as an advanced technology that combines the strengths of deep learning and reinforcement learning, demonstrates significant potential in the field of industrial intelligent control. This paper aims to review the current applications and research progress of DRL in industrial intelligent control. It begins with a systematic introduction to the basic principles and relevant algorithms of DRL. The background of industrial intelligent control is briefly outlined, including common industrial control systems and methods, and an analysis of the application demands and existing challenges of intelligent control in the industry. Subsequently, a detailed review of specific research on DRL in the industrial domain is provided. Furthermore, the paper summarizes the main challenges currently faced by DRL in industrial intelligent control applications, such as data scarcity, sample efficiency, and model generalization. Finally, a conclusion of the current research is presented, along with prospects for future research directions.