Being a porous and composite medium, concrete's microstructure influences its strength, dimensional stability and durability. Adequate characterization of hardened concrete thus is needed for the proper design and maintenance of concrete-based geo-infrastructure such as foundations and dams. Designs and maintenances of concrete structures generally practiced however assume concrete materials to be a homogeneous material or simply a two-phase composite (aggregate and hardened mortar). Within this context, this research aims to develop a method of modeling concrete microstructure using three-dimensional (3-D) finite element analysis and x-ray tomography, such that concrete microstructure and its influence on response to various physical loadings can be accurately represented. For this research, a MATLAB program was developed to read in a set of two-dimensional (2-D) x-ray images taken of a concrete cylinder test specimen, each in the horizontal plane but at 1 mm apart in height. The grayscale intensity value for each pixel in the images gives an indication of the density of the material at that specific point, making it possible to distinguish different material phases, such as aggregate, hydration products of cement, and voids, etc. In this program, pixels are designated to be hydration products of cement, aggregate (fine and coarse), or void based on the grayscale intensity values. With high-resolution X-ray images, finer designations are possible. Each pixel of solid material is converted to a hexahedron finite element with appropriate coordinates for each of the eight nodes such that the elements connect and gaps between the 2-D images are closed. Each phase is then assigned with a set of unique material properties. The 3-D model is imported in ABAQUS for the finite element analysis. A variety of digital tests can be conducted using this technique, which constitutes an economical alternative to experimental tests.