A system identification approach for the analysis of electrochemical noise data is proposed. The most common techniques, used by many investigators, are based on: (i) the ratio of sample standard deviations, which gives no information about the frequency dependence of the electrode impedance, or (ii) Power Spectral Density estimates, which deliver the modulus of the spectrum with large variations at the lowest frequencies. Phase is, almost invariably, not included. In this work, the electrochemical cell is modeled by an input-output model. With the application of system identification techniques, it is possible to identify values of the parameters of the system model. It is shown that this approach delivers a description of the system under study with: smooth electrode impedance curves, and magnitude and phase information. Some results obtained with the most common electrochemical noise analysis techniques are presented for comparison with the proposed approach. A theoretical limitation of the proposed approach appears if a perfect symmetry between both electrodes is considered.