Continuous casting is an important process to solidify molten steel. Efficient and uniform heat removal by water spray without cracking or deforming steel slabs is a major challenge during continuous casting. With the aid of computational fluid dynamics (CFD), the current study presents a numerical model of spray cooling for process understanding and improvement. The Eulerian–Lagrangian approach is utilized to track the movement of water droplets that are injected from a hydraulic nozzle. The initial droplet size is predicted by the linear instability sheet atomization (LISA) model and the droplet distribution satisfies the Rosin-Rammler distribution. Droplet-slab impingement is modeled by the wall jet model and the associated heat transfer accounts for both heat conduction and radiation. Heat transfer coefficient (HTC) on slab surface is validated against a hot-plate benchmark experiment and shows good agreement. The significance of spray standoff distance and spray direction on heat transfer is also investigated. The results indicate the existence of the critical standoff distance beyond which the cooling effect becomes negligible. Spray direction plays an important role in determining the heat transfer rate as well, and insufficient cooling is observed when droplets are sprayed against gravity. Additional 10% to 15% of water will compensate for such low heat transfer. The current study should benefit the steel industry by providing fundamental insights into the process. The current model can also be applied to other types of spray nozzles and can further improve the efficiency of future simulations.