Maize plumpness believed to reflect yield and quality of maize products. Convenient and accurate methods may help identification of maize quality for produces and germplasm resources for breeding. In this study, the 3D reconstruction of maize kernel based on micro-CT technology was introduced to detect anatomical difference between diverse classes of maize kernel. Void spaces measurements constructed for whole maize (Zea mays L.) kernel gives a more accurate volume measurement for density calculations by means of a package of commercial softwares. Moreover, the ratio of cavity and porosity of the entire kernel were calculated based on the 3D CT images. Kernel density, cavities, porosity, and other phenotypic characteristics were closely related to seed plumpness classification. Compared with previous methods, our method significantly improves the calculation accuracy of kernel volume, cavity and porosity, and which is expected to be useful for efficient maize kernel plumpness classification.