The aim of the present study was to accurately assess the changes in the degree of instantaneous pollution of natural river. In this paper, a pollutant diffusion model (LBM-CA) for vegetation open channel is constructed. This model consists partly of Cellular Automaton (CA), and the main task of this part is to simulate the pollutant diffusion. Flow influence coefficient is introduced into Cellular Automata to express the influence of vegetation on pollutant diffusion. In order to obtain the flow influence coefficients for each cellular, Lattice Boltzmann Method (LBM) is introduced in the LBM-CA model to simulate the flow in vegetation open channels. The results show that the LBM-CA model has high accuracy. The simulation results show that the existence of vegetation will lead to the accumulation of pollutants and prolong the residence time of pollutants in the vegetation area. The pollutant limits are embedded in the LBM-CA model to predict the pollution level of the basin at a specific time. LBM-CA model provides a method for simulating pollutants diffusion in natural rivers.