Abstract: The electrical resistivity tomography (ERT) method has become the primary geophysical technique for landslide structure investigation. However, the complex terrain of landslides and the impacts of collapsed surface features render it challenging to arrange survey lines orderly along a straight line in practice. This leads to deviations in the calculations of apparent resistivity. To minimize the impacts of irregular survey lines on the final results, this study developed a scheme for the 3D inversion of measured data. Specifically, this scheme minimized 3D grid subdivision by identifying the rectangle enclosing the minimum area of complete survey lines. Meanwhile, this scheme suppressed the changes in model parameters in the vertical direction by increasing the regularization parameters along the vertical direction of the survey lines. Numerical simulation results indicate that, compared to the traditional 2D inversion scheme, the proposed 3D inversion scheme can significantly improve the identification accuracy of the sliding zone. The measured results of the Baishuihe landslide have verified the effectiveness of the proposed method.