Suppression of image noise is the premise of infrared condition diagnosis of electrical equipment. To improve the effectiveness of suppress white noise in the infrared image of porcelain bushing cable terminal, a kind of adaptive de-noising method based on layer by layer optimal basic wavelet and Bates estimation is put forward in this paper. The method decompose image into R、G and B sub-image. For each one, based on the principle of energy maximum of scale coefficients, the optimal basic wavelet in every layer is selected adaptively. Next, utilizes those optimal basic wavelets to decompose the sub-image, and processing the details of wavelet coefficients by the optimal Bayes estimation criterion. Then, the real sub-image is remained and white noise is removed. Finally, compounds the de-noised R、G and B sub-image together. Simulation results indicate that the method proposed in this paper removes the white noise effectively and keep the information of the original image at the same time. The de-noising ability of this method is better than the sym4 wavelet by having higher SNR and smaller MSE.