Due to influence of atmospheric turbulence and noise contaminated, the object images are always blurred. To observe object in high resolution, combining characteristic of noise, with physical constraint, a novel method of atmospheric turbulence-degraded images blind deconvolution restoration minimization model is proposed. Alternating minimization algorithm based on conjugate gradient method is applied for image restoration. Blurred images by atmospheric turbulence with atmospheric coherent length of 0.1 m for 2 m-diameter telescope and noise with signal-to-noise ratio of 10 dB are restored by the proposed method. The result demonstrates that the drawback of the traditional blind convolution method has been overcome, the influence of atmospheric turbulence and noise has been eliminated and clear observation object images have been restored. The study of blind deconvolution restoration technology purposed is meaningful for ground-based telescope in astronomical observation.