In this paper, a large scale molecular dynamic method for reconstruction of the catalyst layers (CLs) in proton exchange membrane fuel cells is developed as a systematic technique to provide an insight into the self-organized phenomena and the microscopic structure. The proposed Coarse-Grained (CG) method is developed and applied to the step formation process, which follows the preparation of the catalyst-coated membranes (CCMs). The fabrication process is mimicked and evaluated in details with consideration of the interactions of material components at a large scale. By choosing three sizes of the unit box, the relevant configurations of the equilibrium states are compared and analyzed. Furthermore, the primary pores of 2-10 nm in the agglomerates mainly consist of the channel space, which acts as the large networks and could be filled with liquid water. Moreover, various physical parameters are predicted and evaluated for four cases. The active Pt surface areas are also calculated by the current model, and then compared with the experimental data available in the literature. Finally, the pair correlation functions are employed to predict the distributions and hydrophobic properties of the components, providing the information on phase segregation and microscopic structure of the CLs. (C) 2011 The Electrochemical Society. [DOI: 10.1149/2.028203jes] All rights reserved.