This paper proposes a new methodology for matching of objects in video based on the color, shape and region. The objects are segmented and indexed based on the similarity between the frames. The similarity feature such as color, shape and region are measured for the objects between two videos are matched and they are displayed. The similarities between two frames are resulted from three major features such as color, shape and region in order to solve the problem of objects retrieval in video. Applications of our method are the retrieval of objects in video-shot collections or grouping of shots containing the same protagonist into video scenes. MPEG-7 data set are presented and the results are discussed. The experimental results proved the feasibility and effectiveness of the proposed algorithm.