Nowadays, noise reduction in image processing is a major research to enhance image quality. Clustering algorithm is widely used for segmenting the image in image processing. The goal of clustering algorithm is used to make the intrinsic group in a set of unlabeled data. Image acquisition is the first step in image processing. The occurrence of noise during image acquisition can affect the image acquiring result. De-noise based clustering algorithm is classified as de-noise based K-means, Fuzzy C-means and Moving K-means. To improve the image quality, salt and pepper noise is reduced by de-noise based clustering algorithm. During the segmentation process, the effect of salt and pepper noise can be minimized for reducing loss in the images. The experimental result analyses the mean square error and peak signal to noise ratio which gives less mean square error and high peak signal to noise ratio. It shows the image quality without degrading the image quality at the same time unique performance of noise reduction.