Previous studies indicate that particle shape plays an important role in the hydraulic conductivity (k) of granular materials, often represented through the Kozeny-Carman (KC) concept. Several recent studies have improved the accuracy of the KC approach using the particle size distribution (PSD) to estimate the specific surface area of particles but overly simplifying the effect of particle shape. This current study innovatively adopts the Micro-Computed Tomography (CT) technique to compute particle shape parameters of different granular materials (e.g., glass beads, sand and crushed gravel) and then incorporate these parameters into the KC equation to estimate k more accurately, which is then validated with experimental data. The results indicate that k varies significantly according to different particle shapes even if the same mean porosity and PSD are retained. Particles that are less spherical and rounded have a larger fluid-particle contact area (i.e., larger shape factor), hence a smaller hydraulic conductivity. The study suggests a shape factor of 1.28−1.52 for natural sand and 1.84−2.1 for crushed sand and gravel can be used for KC method to estimate k while a porosity-dependent equation is proposed to estimate the tortuosity for different shaped materials.