Known disparity search range is a crucial parameter for many stereo matching algorithms. It is necessary to estimate the disparity search range automatically in real time systems. In this work, we present an approach to estimate disparity search range of each pixel to be matched. A dual-mode camera which can be used to take stereo images or coarse depth maps is developed in our work. Under the constraint of the coarse depth map with the same spatial resolution as the stereo pairs, expected disparity of each pixel is limited within a narrow search range. We demonstrate the effectiveness of the proposed approach on benchmark images with ground truth as well as on images captured in our lab.