Abstract: Since 2000, the Kriging valuation method has gradually become popular in the estimation of mineral resources both domestically and internationally, but the relevant technology is still immature. The main problem is that when technical personnel use the Kriging valuation method for valuation, the selection of key valuation parameters such as the size of the valuation neighborhood, the maximum number of samples, and the size of ore blocks is greatly influenced by subjective human factors, and even has a certain degree of randomness. This may lead to a decrease in the reliability of the valuation results, and sometimes there may be significant differences between the valuation results and the actual grade of the deposit. In order to improve the reliability of Kriging valuation results, this paper intends to use the QKNA method to quantitatively select parameters such as valuation neighborhood size, maximum number of samples, and ore block size through conditional experimental research, so as to provide practical theoretical basis for the selection of valuation parameters and thus improve the reliability of valuation results.