In order to improve the accuracy of the structural dynamic model of the inner gimbal of the electro-optical stabilized sighting system, and to improve the efficiency of the simulation analysis, a structural dynamic model updating method based on the multi-population genetic annealing algorithm was proposed. Taking the inner gimbal of an electro-optical stabilized sighting system as the object of study, we built the Finite Element Model (FEM) by using the dynamic substructure technique. Based on the modal test results, the FEM was updated by using the multi-population genetic annealing algorithm and the correlation analysis to obtain the six-dimensional stiffness value of the inner gimbal bearing in global optimization. An additional experiment was carried out to test the reliability of the updated FEM. The experiment result showed that FEM and the test result were well-matched, the correlation between the modals was over 0. 8, the frequency error was below 9. 05%, and the calculating efficiency was increased by 95%. Thus an effective and feasible model updating method was obtained, which is an important reference to the structure design of the inner gimbal of the electro-optical stabilized sighting system.