The exsisting Landweber iterative algorithms suffer from slow rate of convergence and sensitiveness to noise. In this paper, two common imaging blurs, named motion blur and atmospheric turbulence blur, are discussed about the mechanism of the image blurred, and we propose an improved Landweber iterative algorithm for image restoration. The improved method only expedites the convergence in the signal domain, and inhibits the expansion of noise. The experimental results demonstrate that the proposed method can still improve the restoration accuracy of results at the same time of speeding up convergences, and further show the application effect by applying the method to the remote sensing image and high-speed railway image.