Tracking is to match the correspondences between a frame, and objects which are in motion. It is usually performed in the context of higher-level applications that require the location and / or shape of the object in every frame. Typically, assumptions are made to constrain the tracking problem in the context of a particular application. In this i had categorizes the tracking methods on the basis of the object and motion representations. In all the previous works, tracking is only depends on the low level correspondences between frames, but here I had proposed that a tracking is done based on both low and high level correspondences, and the tracked result is feed back into a recognition part to find category of an object it is the final result. Tracking is done based on the detection of the movable objects, it’s done by Gaussian Mixture Model. Harris Corner is used for tracking and the classification is based on Naive-Bayes Classifier. This paper is works well in a challenging areas like drastic view change, background clutter, and morphable objects, crowd, complex scenes.