Abstract: This paper addresses the issue of center carbon segregation during the continuous casting process of 70 steel and proposes a Genetic Algorithm-guided Ant Colony System(GA-guided ACO). The Genetic Algorithm is used to pre-generate the initial pheromone distribution, and a dynamic hybrid update mechanism incorporating the genetic algorithm is introduced. This effectively solves the premature convergence problem commonly encountered in traditional Ant Colony Optimization(ACO) algorithms in multi-objective optimization. In the later stages of the iteration, the ACO continues to leverage its powerful global search ability. Compared to the traditional ACO, the convergence speed is improved by 53.13%. In this work, a coupled solidification and heat transfer model was implemented, which outputs key parameters such as temperature gradient and solid fraction, providing the basis for the objective function calculation for the algorithm. This forms a closed-loop control framework of "model prediction-algorithm optimization". The industrial applications show that the optimized secondary cooling process reduces the center segregation index by 5.8% at a casting speed of 2.2 m/min. The segregation index is reduced by 5.38% and the total water consumption is decreased by 4.63%, when the casting speed is 2.4 m/min. The 70 steel billet segregation quality is improved and production cost is reduced. The experimental results demonstrate that the improved algorithm provides an effective solution for the optimization of the segregation of 70 steel continuous casting process.