In this paper, a fast blind restoration algorithm for motion blurred image was proposed, using a robust algorithm based on Radon transform-domain to determine the blur kernel function, then a modified total variation algorithm was used to restore the blurred images. Its cost function is the sum of three terms corresponding to total variation l2-norm regularization, least squares fidelity term and l1-norm fidelity term. Compared with Fergus' and Levin' algorithm, the experiment results show that the algorithm for a class of motion blurred image caused by the linear movement parallel to the lens has higher speed and good recovery effect.