Abstract: In order to reduce the influence of large variation of lake bottom elevation sampling points on interpolation calculation and obtain high precision lake bottom topography, the influence of variation degree of elevation sampling points on interpolation is introduced into the ordinary Inverse Distance Weight (IDW) interpolation, and a lake terrain partition interpolation method considering the variation degree of elevation sampling points is proposed. By introducing the optimization method of orthogonal test optimization, the optimal power value of inverse distance in each partition is obtained. Taking Taihu Lake as an example, the interpolation effects of several interpolation methods on the bottom topography of Taihu Lake are compared. The results suggest that the regional inverse distance weighted interpolation method keeps good adaptability, it is verified by the measured elevation value, and the root mean square error is the smallest. Considering the terms of storage capacity error, comparing with Kriging method, inverse distance weighted interpolation method as well as natural neighborhood method, the average relative error is decreased by 0.97%, 0.90% and 1.37%, respectively. The interpolation effect is obviously superior to these interpolation methods. Therefore, when it comes to large topographic relief of the lake, the subregional inverse distance weighted interpolation method can obtain a high-precision lake bottom shape, and is of high application value for different types of lake bottom topography interpolation calculation.