Image is a collection of row and column that is called pixel values. Extracting best matched image from large collection of database is emerging task. Image retrieval is mainly used in image processing, pattern recognition and computer vision. CBIR technique used in many areas such as medical, academic, art, fashion, entertainment. Generally image have colour, texture, shape and size are relevant feature so extract all the relevant and irrelevant features of image. Colour histogram of image formed by divided graph into levels and mean and standard deviation are calculated by pixel value of each level. After extracting all the feature of image applies SVM i.e. supervised learning algorithm get optimal result for image classification.