The paper deals with the use of the stochastic approximation method for estimation parameters of the fluidization process and detecting a change of the process mode based on experimental photographs of local part. The fluidization process is treated as a process of diffusion type, described by the diffusive equation. Because the solid particles are essentially discrete ones and the moments of process observation are also discrete, the diffusive equation is treated in the difference form. The problem of modeling of the fluidized bed is reduced to the estimation of coefficients of difference equations system. The method of stochastic approximation is used for this estimation. The method allows identifying the regime of the fluidization process in its dynamics and estimating the parameters of the process. The approach and the estimating algorithm were verified using the especial experiments.