Abstract: This paper proposes a video object segmentation approach based on dynamic bilateral grid that aims to handle the problems such as inseparable color,object occlusion and data redundancy.Firstly,a high dimensional grid is initialized for the segmentation.Then we compute the data cost for each node in the high dimensional grid by accumulating the effects from all the labeled pixels.After that,the grid can be segmented by the well-known Graphcut algorithm given the data cost and smooth cost.Last,the final mask of each video image is retrieved from the high dimensional grid.In experiments,the proposed approach is evaluated by different measures,and the results on DAVIS benchmark demonstrate the effectiveness and robustness in comparison with other state-of-the-art algorithms.