Blind deconvolution of adaptive optics images plays an important role in the area of space object surveillance and astronomical imaging. This paper proposes a high resolution blind restoration method for adaptive optics images based on wavelength diversity. Firstly, we utilize Zernike polynomial to establish the relation between point extension functions of the two channel images under the frame of Fourier optics theory. Secondly, we derived the general cost function of the difference between the object image and the collected images by using of the least square principle. The point spread functions of each image could be estimated by minimizing the cost function, and the object image could be obtained subsequently. The algorithm proposed in this paper can fundamentally improve the ill-posed problem of blind deconvolution. Experimental results show that our method is robust and can effectively improve the imaging quality of turbulence degraded images.