<p>This study presents a parameter identification framework for multibody dynamic systems based on the Enhanced Response Sensitivity Approach. The identification problem is formulated as a time-domain nonlinear weighted least-squares problem using either Newton-Euler or Lagrangian dynamics, depending on the sensing modality and constraint structure of the system. To improve numerical robustness in the presence of measurement noise, strong nonlinearities, and parameter coupling, Tikhonov regularization and a trust-region strategy are incorporated into the iterative solution procedure. Two representative benchmarks, namely a planar two-link mechanism and a PUMA560 manipulator, are used to assess the performance of the proposed framework. The results show that the method can accurately estimate key physical parameters, including mass, inertia, and friction-related quantities, under noisy measurements and different modeling settings. A comparison with a standard Levenberg-Marquardt solver is further included to provide a quantitative point of reference. The study indicates that ERSA offers a structured and effective approach for parameter identification in multibody systems.</p>