Digital images play a significant role both in daily life as well as in areas of research and technology. Data sets collected by image sensors are generally infected by noise. Imperfect instruments, problems with the data acquisition process, and interfering natural phenomena can all degrade the data of interest. Furthermore, noise can be introduced by transmission errors and compression. Thus, denoising process is often a necessary and the first step to being taken before the images data is being analyzed. It is necessary to apply an efficient denoising technique to compensate for such data corruption. Image denoising still remains a challenge for researchers because noise removal introduces artifacts and causes blurring of the images. The challenge of any image denoising algorithm is to suppress noise while producing sharp images without loss of finer details. A modified method based on patch-based image modeling is proposed in this research work. The main part of proposed method is the use of image nonlocal self-similarity (NSS) prior. NSS prior refers to the fact that a local patch often has many nonlocal similar patches to it across the image. This fact significantly enhances the denoising performance. Patch Groups are extracted from training images by putting nonlocal similar patches into groups. According to these Patch Groups, Gaussian Mixture Model learning algorithm is developed to learn the NSS prior. The whole process is repeated 4 times to make the system learn more and more. The iteration process regulates and optimized some of the variables. MSE (Mean Square Error), PSNR (Peak Signal to Noise Ratio) and Correlation coefficient has been taken as output parameters to evaluate the performance of proposed system. MATLAB R2013a has been taken as implementation platform using image processing toolbox.