AbstractAn advanced cellular automaton model (CA) was designed for earthquake preparation process in a single fault in this paper. By designing the energy exchange between the circumstance and the model by loading and the nonlinear interaction among the cells of the model, we tried to understand the mechanism of seismicity feature. Compared with the cellular automaton models published before, the cellular automaton model in this paper is much closer to the natural seismic preparation system because of the improvement of sampling and the introduction of time‐dependent frictional mechanism. The results show that the method of sampling is important for the temporal heterogeneous feature of the man‐made seismic catalog and the breaking of each cell and that the events with low magnitude and large magnitude obey respectively remarkably different magnitude‐frequency relationships.