Aiming at the problem of establishing the mathematical model of the electromagnetic-driven micro-electro-mechanical system (MEMS) micromirror applied to the laser radar, the discrete model of the electromagnetic-driven MEMS micromirror is established by combining the mechanism analysis method with the input-output method. A recursive least squares method with variable forgetting factor is proposed to identify the model parameters of electromagnetic-driven MEMS micromirror. By making the forgetting factor dynamic, the problem of "data saturation" is solved, so that as much input and output data as possible can play a role in parameter identification, and the accuracy of parameter identification is improved. Through the simulation and experimental verification of this method, the results show that the error of the model obtained by recursive least squares identification with variable forgetting factor is reduced by 9.2% compared with traditional recursive least squares identification.