针对普通Kriging插值方法中变差函数模型参数求解难、模型选择具有主观性、并不一定能选出最优变差函数模型的局限性,将最小二乘支持向量机与Kriging方法结合起来,利用最小二乘支持向量机较强的非线性回归能力从实际数据中重构变差函数。为了评价该方法的插值精度,将它和传统普通Kriging插值方法相比较,并利用研究区实际煤层厚度数据进行了验证。 In order to solve the parameter of variogram model in ordinary Kriging interpolation method, the model selection is subjective and the limitation of optimal variogram model can not be selected. The least squares support vector machine is combined with Kriging method. Least square support vector machine strong nonlinear regression ability to reconstruct the variogram from the actual data. In order to evaluate the interpolation accuracy of this method, it is compared with the traditional Kriging interpolation method and verified by the actual coal thickness data in the study area.