<p>The identification performance of a typical nonlinear mechanical system, as a gantry equipment, is affected not only by the most appropriate method to be applied. The excitation signal used plays a crucial role in conjunction with the method. In this work, the classical Cart-Pendulum nonlinear system is simulated with different input excitation to provide the dataset required for the tasks of System identification. Gaussian White Noise (GWN) and Pseudo Random Binary Sequence (PRBS) signals are compared as excitations for evaluating the identification effectiveness when applying the Autoregressive Moving Average with External input (ARMAX) method. The choice of a traditional algorithm becomes appropriate in the context of this work since the focus was to analyze the influence of excitations on the identification performance. The identification process was also submitted to an additive measurement noise of 3% during the system simulations, which fulfilled an amount of 100 datasets for each identification setting. Based on a statistical analysis of 100 identifications performed for each configuration, the results showed an identification effectiveness of more than 96%. That outcome was achieved specifically when the ARMAX was applied in conjunction with the PRBS adjusted with a clock frequency reduction compared to GWN excitation. In addition to the ARMAX method, the comparative tests with other identification algorithms for both the GWN and PRBS excitations indicated a feasible and leaner possibility of identification applying the ARMAX when the nonlinear system is excited by a PRBS appropriately configured in terms of switching determined by the clock frequency.</p>