Abstract: Based on the regular analytical data of the plant, the feedstocks of Deep Catalytic Cracking (DCC) on a molecular scale were simulated by the Monte Carlo method combined with Structural Oriented Lumping. The results show that Monte Carlo method can simulate the characters of the DCC feedstock on a molecular scale very well, and the predicted values of feedstock properties such as average molecular weight, saturated hydrocarbon content, aromatic content, resin content, asphaltene content, carbon content, hydrogen content, sulfur content and nitrogen content are in good agreement with the plant data during two operation states. Moreover, the molecular matrices formed in the process of Monte Carlo Simulation lay a foundation for modeling the reaction kinetics of DCC on a molecular scale.