Abstract: A novel image processing approach is proposed to reduce the mixed noise in surface mount technology soldering image based on wavelet packet transformed adaptive threshold. At first,by using wavelet packet transform,the approach not only decomposes the image into the low frequency part but also into the high frequency part of image in several scales. After analyzing the wavelet packet tree coefficients,they were processed with the Wiener filter,and kept the wavelet packet tree low frequency coefficients without change. Secondly, an improved wavelet adaptive threshold algorithm is proposed to denoise the mixed noise again. At last,the inverse wavelet packet transform is applied to reconstruct the image and median filter is used to smooth the image. The experimental results have illustrated that the approach can obtain a better result in soldering image denoising compared with the conventional methods and can retain the image edges very well.