Parametric models for the representation of the power spectrum and bispectrum of a bank of data by the use of a two-stage, auto/cross-correlation matching (ACM) technique are developed. First, the applicability of the ACM technique for ordinary spectral modeling is discussed. Then, the parametric bispectral modeling technique is formulated using third-order cumulant information. This autoregressive moving average (ARMA) technique allows for the simultaneous determination of both the AR and MA filter coefficients. This procedure can expedite the analysis of phase coupling, system nonlinearities, and other features which cannot be captured by ordinary spectral analysis. Techniques are demonstrated numerically through the analysis of Space Shuttle flight data.