Abstract: Some image restoration algorithms based on complex wavelets are discussed, and HMT(hidden Mar-kov tree) model is applied to the ForWaRD image restoration algorithm. Firstly, we use the Wiener filter to deblur the observed image in the Fourier domain, and then remove the unwanted noises in wavelet domain. Simulations are done for solving the typical convolution and noised linear degraded modeling. By comparing the results based on complex wavelets with those of the real orthogonal wavelets, we find that the former method performs better by visual feelings and objective criterions, especially the method of HMT ForWaRD based on complex wavelets.