从钻孔中建立三维地质模型对于提供对复杂地下地质的连贯理解至关重要，但传统克里金插值方法在地质建模中受限于变异函数模型选择的主观性和形式固定性，难以精准反映空间数据的波动特征。为此，提出了通过支持向量机构建实验性半变异函数与空间距离的非线性映射关系，同时利用鲸鱼优化算法优化支持向量机的关键参数来提高拟合精度的克里金插值算法。在安徽牛山西方解石矿区的案例研究中，构建了方解石矿的三维地质模型，验证了改进算法的有效性。实验中选取均方误差作为算法性能指标，与优化前的Kriging插值法相比，优化后的克里金方法在ZK602、ZK1001、ZK1002验证钻孔的均方根误差从2.06降低至1.25，下降幅度达38%，且对方解石矿层的插值结果误差分别为0.10、0.31、0.50 m。改进方法的插值结果更接近实际值，为复杂地质环境中的三维地质建模提供了一种高效、可靠的技术方案。

Establishing a three-dimensional geological model from a borehole is crucial for providing a coherent understanding of complex underground geology, but traditional Kriging interpolation methods are limited by the subjectivity and fixed form of the selection of the variation function model in geological modeling, making it difficult to accurately reflect the fluctuation characteristics of spatial data. Therefore, a nonlinear mapping relationship between experimental semi variogram function and spatial distance was proposed through support vector mechanism, and a Kriging interpolation algorithm was used to improve fitting accuracy by optimizing key parameters of support vector machine using whale optimization algorithm. In a case study of the Niushan west calcite deposit in Anhui Province, a three-dimensional geological model of the calcite mine was constructed to verify the effectiveness of the improved algorithm. In the experiment, the mean square error was selected as the algorithm performance indicator. Compared with the pre optimized Kriging interpolation method, the optimized Kriging method reduced the root mean square error of the verification boreholes in ZK602, ZK1001, and ZK1002 from 2.06 to 1.25, a decrease of 38%. The interpolation results of the calcite ore layer showed errors of 0.10, 0.31, and 0.50 m, respectively. The interpolation results of the improved method are closer to the actual values, providing an efficient and reliable technical solution for 3D geological modeling in complex geological environments.