In the process of BOF steelmaking, in order to figure out and optimize the amount of alloy addition during deoxidizing and alloying of tapping, alloy addition multiobjective optimization model is developed. Firstly, the prediction model of ladle element content is developed with support vector machine(SVM). One objective function of multiobjective problem is the error between the regulations and the result of the prediction model of ladle element content, and the other objective function is the cost of alloy addition. Then the result of alloy addition calculation is made as initial value of the optimization model, and the alloy addition is optimized by using modified dynamic neighborhood multiobjective optimization-particle swarm optimization(MO-PSO) so as to determine the Pareto optimal solutions effectively. The final amount of alloy addition is decided according to the practical requirement. Finally, experimental results show the effectiveness of the proposed model.