Marching Cubes (MC) algorithm is a typical representative for surface reconstruction method.However there are still some shortcomings about the represent accuracy of the isosurface which extracted from sparse volume. An adaptive algorithm of Marching Cubes based on volume warping is given. The proposed algorithm uses an iterative process to adaptively displaces the vertices of the cubes. After each iteration

the cubes obtain more accurate representations of the isosurface

and the difference between the implicit and the explicit representations is reduced. Region growing strategy is used to segment volume which avoid the calculation for all vertices. And parallel approach is used to optimize the proposed algorithm which promote algorithm-s efficiency. The experiment results show that this algorithm enhance the accuracy and ensure the instantaneity and interactivity of 3D reconstruction for sparse volume.