While it is now possible to image the three-dimensional structure of soil using high-resolution tomography, none of the techniques can simultaneously image the distribution of the resident soil microbes. This means that it is not possible to visualize soil microbes in their habitat. Consequently, the impact of soil structure on microbially mediated processes cannot be reliably modeled. Biological thin sections offer the opportunity to simultaneously image microbes in structure but are necessarily restricted to two dimensions. Therefore a methodology is required to simulate three-dimensional structures from two-dimensional thin sections of soil that is extendable to simulate the spatial distribution of a range of soil components. We developed a model that is capable of using data gathered from two-dimensional sections to predict the three-dimensional structure of soil. An object-oriented approach to modeling was used to allow the individual representation of each structure voxel. This allows the model to encapsulate both data, presented here, and the subsequent addition of components such as microbial distribution and related diffusion–respiration processes together in a three-dimensional lattice of voxels. The model was validated using data derived from three-dimensional x-ray tomography images of soil structure, and using two-dimensional sections through that data set to predict three-dimensional structure. A range of metrics was used to compare the modeled and imaged three-dimensional structures. The comparison shows that the metrics for the modeled structures agree with those derived from the three-dimensional images for higher porosities, but that systematic differences occur for the lowest porosity soils (<11%). This is due to problems relating to the prediction of rare events such as the presence of large connected pores in low-porosity samples.