A novel methodology of maximum information entropy theory combined with structure oriented lumping (SOL) and Monte Carlo (MC) method was developed to simulate the feedstock for secondary reactions of fluid catalytic cracking (FCC) gasoline at the molecular level. The SOL method was applied to represent the feedstock configuration framework; a molecular library consisting of a large ensemble of computational molecules was created by stochastically assembling structural increments using MC method; then, the maximum information entropy method (MIEM) was introduced to adjust the mole fractions of the structural increments in the molecular library to achieve a closer matching with the actual analytical characteristics. Three samples of catalytic cracking gasoline from the industrial FCC units of China were used to validate and evaluate the proposed method. The simulation results of the feedstock properties, such as the average molecular weight and the weight fractions of main components including paraffin, olefins, naphthenes and aromatics agreed well with the experimental data. (C) 2015 Elsevier B.V. All rights reserved.