基于最小描述长度（ＭＤＬ）准则，提出了一种新的自回归滑动平均（ＡＲＭＡ）模型结构辨识方法．该方法将 ＡＲＭＡ模型的结构辨识分两步进行：首先利用超定辅助变量乘积矩（ＯＩＶＰＭ）的最小特征值确定自回归（ＡＲ）部分的阶，然后利用协方差矩阵的特征值估计滑动平均（ＭＡ）部分的阶．方法的可行性与有效性通过大量的数值仿真得到验证． Based on the minimum description length (MDL) criterion, a new autoregressive moving average (ARMA) model structure identification method is proposed. This method identifies the structure of the ARMA model in two steps: First, the order of the autoregressive (AR) part is determined by using the minimum eigenvalue of the OIVPM (overriding auxiliary variable), and then the moving average is estimated by using the eigenvalues ​​of the covariance matrix MA) part of the order. The feasibility and effectiveness of the method are verified by a large number of numerical simulations.