It is well known that soil structure, especially its pore networks, affects many fundamental soil properties. However, it is still challenging to quantify and model soil structure, particularly with regard to its impact on dynamic processes such as water flow and chemical transport. Improved Pore-Solid Fractal (PSF) model and multi-point geostatistics are used in this study to characterize the soil structure of an agriculturally important soil in Pennsylvania (Hagerstown silt loam) based on the spatially exhaustive data collected with Micro Computing tomography (Micro-CT). Micro-CT provides a nondestructive means of imaging intact soil structure at high resolution (around 100µmx100µmx100µm) and allows simultaneous real-time flow and transport measurements. We investigate large undisturbed soil columns (0.1 m in diameter and 0.3 m in length) of the well-structured Hagerstown soil. Improved Pore-Solid Fractal (PSF) model has the capability to take into account the pore size distribution, pore-solid interface, pore connectivity, and spatial correlation of pore network. Traditionally, two-point correlation function is used to describe the correlation of the porous media but it can not represent the geometries of the objects. Multi-point geostatistics overcomes this shortcoming and is able to consider both spatial correlation and the shape of soil structure by parameterizing the pore geometry within training image. Therefore, multi-point geostatistics has a promising potential to simulate the spatial distribution of soil structure, especially the macropore network. The comparison of the PSF model and multi-point geostatistics is expected to shed light on soil structure's impacts on complex physical, chemical and biological processes occurred in the soil.