Abstract: For solving the incorrect stereo matching problem of visual images, an adaptive windows-constrained stereo matching algorithm based on the dynamic time warping (DTW) is proposed in the paper. In this algorithm, we first find out a pair of compatible sequences from images according to the constraint epipolar. In searching the sequences, we use the technique of gray correlation in adaptive windows to obtain two matching sequences which may not have to be in the same lengths. Then, the dynamic programming (DP) method with DTW is used to find out an optimum matching path under the continuity constraint. Detailed disparity map can be obtained from the matching paths and the coordinates of the matching sequences in images. Experiments show that this algorithm is highly efficient in computation, and generates nice result of stereo matching.