Abstract: Equipment and mathematical models of a hot strip mill laminar cooling control system were described. Mathematical models of the system consist of models for air cooling, laminar cooling, feedback control, and self-learning. The original laminar cooling system was unable to effectively handle the relatively large changes in strip speed in the cooling area, thereby significantly compromising coiling temperature control accuracy. Speed feedforward compensation was proposed to achieve more adequate coiling temperature control. Improved speed feedforward compensation methods are provided in the process automation system and in the basic automation system. The optimized laminar cooling system was stable and reliable with high control precision, and metallurgical properties of the product were improved significantly, thereby laying a foundation for users to develop new products.