To understand the mechanism of self-assembly and to predict the evolutionary pattern of the fusion-fission system over a long period of time, studying the dynamics of these processes is of great significance. The trajectories from molecular dynamics (MD) simulations of self-assembly processes contain numerous latent fusion and fission events. To analyze the fusion and fission events from the simulated trajectory, in this article, a dynamic clustering approach was developed by comparing the changes of monomer composition within clusters over simulated time. The rates of fusion and fission events obtained from dynamic clustering analysis were further coupled with the population balance model (PBM), and the evolution of the molecular assemblies calculated is reasonably consistent with the corresponding MD simulation results. This approach provides a new idea to analyze the big data of molecular self-assembly dynamics obtained from MD simulations, and it offers a computationally achievable approach to connect discrete events at the molecular scale with continuum equations such as the PBM at the macroscale.