The paper presents an efficient algorithm for deblurring natural images using a variational Bayesian approach to estimate the point spread function (PSF) and a modified Lucy Richardson algorithm for deconvolution. It highlights the challenges of restoring blurred images, particularly in real-world scenarios where factors like camera motion and environmental conditions contribute to image degradation. Experimental results demonstrate the algorithm's effectiveness compared to existing methods, particularly in synthetic scenarios, although limitations remain for complex natural blurs.