This study investigates the use of Monte Carlo simulation for the determination of release properties from cubic inert matrices. Specifically, the study has focused on factors including porosity, surface area and tortuosity. The release platform was formed by simulating matrices with different ratios of drug and excipient, which undergo drug release in a uni-directional (two-face) or omni-directional (six-face) process. Upon completion of each simulation the matrix 'carcass' was examined and porosity and tortuosity of the medium evaluated. The tortuosity of the medium was evaluated directly by a blind random walk algorithm. These parameters as well as the release profile were then studied with respect to common mathematical models describing drug diffusion (the square-root, power and Weibull models). It was found that, depending on their composition, the matrices systems were either homogeneous or heterogeneous in nature. Furthermore, it was found that the physical parameters could be successfully fitted to the a and b constants of the Weibull model. This approach allows the prediction of drug release from an inert matrix system with the knowledge of a few physical parameters.