Large-scale integrations of power-electronics devices have introduced the stability challenges to the conventional power system. The stability of the power-electronics-based power systems, which are modeled by a Multi-Input Multi-Output (MIMO) transfer function matrix, can be analyzed based on the Nyquist Criterion. However, since no or limited information about the internal control details, this matrix can only be obtained using the measured data. On the other hand, the elements of the matrix will change along with the operating point of each power-electronics converter, which introduces the challenge to guarantee the interaction stability of each inverter at different operating points. In this paper, a data-driven method is proposed to overcome this operating-point dependent challenge. An artificial neural network (ANN) is used to characterize the operating-point dependent model of power-electronics-based power systems. The comparison results confirm the accuracy of the impedance model obtained by this data-driven modeling method.