Abstract: On account of long-time ore blending with multi elements & ore dropping points， and rough ore blending upon the uneven grade distribution of blast heaps， so an open pit automation ore blending method based on goal programming was put forward. Blasting heap was discretized into a number of cell block， then based on the sampling data of rock powder， the accurate prediction the grade of blasting heaps was realized by use of the geostatistics method. Considering the priority of ore blending and taking the minimum grade deviation as the objective function， the optimization model of the open pit ore blending based on goal programming was established. Using C++ language programming and solving the model by the LPSlove， then output the open pit ore blending optimization scheme and enclose the areas of mining area automatically. The method is applied to a large polymetallic open pit in Inner Mongolia， resulting in a quick ore blending computation time， a good ore blending result within the range of allowable error， and a high-efficiency of ore blending.

Key words: multielement, ore blending, accurate prediction, goal programming, automatic enclosing