In the paper, two nonlinear estimation methods based on nonlinear chaotic theory, surrogate data method and Lyapunov exponents, are used to distinguish the difference of nonstationary signals. After brief introduction of the corresponding algorithms, two typical different nonstationary signals measured from a thinplate structure with different nonlinear constraining boundaries are taken to compare by using the above two methods respectively. The obtained results demonstrate that the apparently similar signals are distinguished effectively in quantitative way with applying above nonlinear chaotic analyses.