Abstract: Addressing the issue of inclusion defects in automotive panels caused by slag entrainment in molten steel due to coriolis forces during the final stage of continuous casting ladle pouring, this paper systematically investigates the dynamic behavior and influencing factors of slag entrainment in the molten steel during the ladle pouring process by establishing a numerical simulation method based on the k-ε turbulence model coupled with the VOF multiphase flow model. The research reveals that the slag entrainment process can be divided into four stages: initial depression, formation of conical vortices, turbulence entrainment, and explosive slag entrainment. Among these, the steel throughput is the core factor affecting the critical slag entrainment level, showing a significant positive correlation with the critical level. Factors such as ladle age, slag layer thickness, and steel slag viscosity have negligible effects. Based on this, a prediction model for critical remaining steel in the ladle is constructed, incorporating multiple parameters such as steel throughput, casting speed, and ladle age. Additionally, a calculation equation for critical remaining steel is proposed, with a prediction error of less than 3%. Furthermore, an intelligent remaining steel control and early warning system is developed, enabling precise early warning before slag entrainment through real-time collection of process parameters. Industrial applications demonstrate that this system reduces the amount of remaining steel in the ladle by over 2 tons per ladle, lowers the slag entrainment rate to 0.5%, and decreases the incidence of inclusion defects by 0.1%. This significantly enhances the purity and surface quality of automotive panel steel, while reducing production costs and achieving significant economic benefits.