A voxel-based method of recognizing and visualizing the potential water leakage gaps of automobile assembly was proposed for finding the water leakage problem during the new car model design phase. Based on the voxelization of the automobile assembly, the boundary voxels representing the position of water leakage were determined by defining the gap voxels among neighboring parts, reconstructing triangular meshed surfaces and extracting the boundary edges of parts. The gap distances were calculated and the gap surfaces were constructed based on the presented algorithm of mapping points generation. The errors from the boundary edges of gap surfaces to part surfaces were reduced by the self-adaptive subdivision method controlled by the chordal height error, and the gap surfaces were visualized by the color map. The method can be used to precisely identify all the potential positions of water leakage and visually display the gap sizes, which provides a possibility for quickly verifying the reasonability of water tightness design, and thus avoiding the reputational and financial losses caused by the emergence of congenital design defects. The method can also be applied to the design of large-scale and complex products in other industries such as aircraft, ship and chemical engineering.