A new image denoising method of wavelet shrinkage threshold based on Bayesian estimation was proposed. The coefficients of wavelet transform of image were estimated by minimizing Bayes risk in the proposed method. The estimation is not only related to the orientation and the level of the subband, but also to the wavelet coefficients. The experimental results show that the denoising effect of the proposed method is better than that of other methods based on wavelet shrinkage. Although the denoising effect of this method is worse than one of Wiener filters at low peak-signal-noise ratio, it is better than Wiener filters at high peak-signal-noise ratio.