Today, in the age of images and digitization relevant retrieval is quite a topic of research. In past era, the database was having only text or database was low dimensional type. But with every new day thousands of pictures are getting added into the database making it a high dimensional data set. Therefore, from a high dimensional dataset to get a set of relevant images is quite a cumbersome task. Number of approaches for getting relevant retrieval is defined, some includes retrievals only on the basis of color, while some include more than one primitive feature to retrieve the relevant image such as color, shape and texture. In this paper experiment has been performed on the trademark images. Trademark is a very important asset for any organization and increasing trademark images have developed a quick need to organize these images. This paper includes the implementation of HSV model for fast retrieval. Which use color and texture so as to extract feature vector. Experiment takes query image and retrieve twelve most relevant images to the user. Further for performance evaluation parameter used is Precision and Recall.