The purpose of this research was to evaluate and optimize the performance of a moving bed granular filter (MBGF). A moving bed granular filter was constructed to test the removal of char particles from a high temperature product stream produced by a biomass pyrolysis system. Tests described in this thesis were performed under cold flow conditions. Cold flow testing was designed to identify important physical parameters of the system that characterize filtration efficiency. Parameters such as gas velocity, bed depth, downcomer diameter, and granule size as well as dust characteristics were investigated. A correlation of bed depth and gas velocity has also been developed to allow design engineers to design a MBGF system based on known gas flow and dust conditions. Experiments performed under cold flow testing with coal-derived fly ash verify levels of filtration efficiency exceeding 99% for some filter arrangements. However, char particles from the pyrolysis system do not filter as efficiently as fly ash due to different particle properties such as particle size, surface roughness, shape, and density. Particles with high kinetic energy experience rebound from the granular surface, and consequently have low filtration efficiencies. Filtration efficiencies are high for particles that do not experience bounce or high drag forces.