Abstract: Cut−off grade is an important decision parameter in mining operations. In multi metal mining projects, determining a reasonable cut−off grade is the foundation for obtaining greater economic benefits for subsequent mining. Based on the maximum net present value method, a dynamic optimization model for the comprehensive cut−off grade of a polymetallic mine based on ICA algorithm was constructed for the two−stage production process of mining and selection. The optimal cut−off grade of a certain silver lead mine was dynamically determined. The example application shows that the model is suitable for determining the cut−off grade of polymetallic mines. During the lifespan of the mine, the optimal cut−off grade for a certain silver lead polymetallic material determined by the ICA algorithm was 2.619%, which later decreased to 1.331%. The total net present value of the mine was 1274.5753 million yuan; Compared with the Lane method, this model has global search ability and is more advantageous in dynamic optimization of cut−off grade indicators in the later stage of mining, providing new ideas for determining reasonable cut−off grade indicators in mines.