Abstract: In this paper, the structural parameter identification and the load identification of inverse analysis are studied. The resultant nonlinear global optimization problem is solved using a hybrid genetic algorithm in which the BFGS algorithm is chosen for local search. With no information of input excitation forces, a calculation strategy of decomposition inversion is presented, which improves the robustness and convergence rate of the hybrid genetic algorithm. Numerical simulation studies are performed. For verification purpose, both the noise-free and noise-included output responses are considered. Results show that the proposed method is efficient not only for structural parameter identification but also for load identification in inverse analysis.