Image denoising is an effective way to enhance the quality of scanning electron microscope (SEM) images, and improve the accuracy of detection analysis and automatic operation. However, most existing denoising methods rely on global parameters and are not suitable for noise removal in all regions of the image. To address this issue, this paper proposes a denoising method based on the adaptive anisotropic partial differential equations (AAPDE), which can adaptively select denoising parameters for different regions according to the gradient amplitude of SEM images and effectively solve the problem of global parameters not being applicable to noise in all regions of the image. Both simulation and experimental results demonstrate that the proposed method outperforms global denoising methods in terms of denoising performance.