Burdening calculation is the important preparatory process for special aluminum smelting, which affects production efficiency and product quality directly. To improve production quality and burdening efficiency and reduce the costs of material and storage logistics, burdening optimization model of special aluminum alloy is built with melting loss of chemical elements and cyclic utilization of waste products. For multi-objective and nonlinear characteristics of this model, real code rule is designed that the amount of raw materials and charging time are decision variables, and the multi-objective algorithm based on NSGA-III are proposed to solve it with distributed estimation strategy. The model and algorithm are verified by means of the simulation experiment with real production data. The experimental result shows this optimization model and the proposed algorithm meets the requirements of enterprises and can solve the burdening optimization problem for special aluminum alloy. The proposed algorithm can obtain better solutions compared to NSGA-III when solving the model.