A new model for estimating macroscopic permittivity was proposed to predict filler particles' dispersion states in a particulate composite material. In the model, the estimation targets are randomly packed composite materials. The composite materials were represented as a cluster of unit cells. A proposed layer structure model connected the unit cells. The macroscopic permittivity was estimated by calculating the synthetic capacity of the cluster. The proposed model was validated by comparisons between estimated and measured macroscopic permittivity of several particulate composite materials. It was also identified that the proposed model could estimate the permittivity more accurately than an existing theoretical equation's one due to considering the effects for the dispersion states of filler particles. Furthermore, it was indicated that the proposed model could also estimate the dispersion states of filler particles by the measured permittivity. The applicability of the method was confirmed by comparisons between estimated and experimental dispersion states of filler particles. (c) 2021 The Society of Powder Technology Japan. Published by Elsevier B.V. and The Society of Powder Technology Japan. This is an open access article under the CC BY-NC-ND license (http://creativecommons. org/licenses/by-nc-nd/4.0/).