A separated model parameter identification method for electro-optical stabilized platform servo systems was presented. The model of servo systems was separated into a linear part and a nonlinear part. In the process of parameter identification of the linear model, a differential inverse M sequence signal was utilized to excited servo systems, and least square method was used to identify the parameters of the linear discrete model part to reduce the influence of the nonlinear dynamics. In the process of parameter identification of the nonlinear model, the coupling dynamics between friction and unbalanced mass was considered, and the static parameters of the LuGre model and the parameters of the unbalanced mass model were firstly identified by using the least square method; Then, the dynamic parameters of the LuGre model were obtained by friction measurement through a current signal. An experimental system for single-axis electro-optical stabilized platform was constructed, and the dynamic model parameters were indentifies by proposed method. Experimental results indicate that the identification value of unbalanced mass torque is 0.183 N·m, which is a little larger than the theoretic value 0.18 N·m, and satisfied with the identification precision. So the parameters of linear and nonlinear models can be obtained by the presented identification method.