The traditional color-based mean-shift tracking algorithm is popular among tracking methods due to its simple and efficient procedure, however, the lack of dynamism in its target model makes it unsuitable for tracking objects which have changes in their sizes and shapes. In this paper, we propose a fast novel threephase colored object tracker algorithm based on mean shift idea while utilizing adaptive model. The proposed method can improve the mentioned weaknesses of the original mean-shift algorithm. The experimental results show that the new method is feasible, robust and has acceptable speed in comparison with other algorithms.15 page,