Adaptive ground moving target indication (GMTI) algorithms based on the sample matrix inversion require the availability of a secondary data (training data) set to determine the adaptive filter. A polarization-aided GMTI method is devised in this letter for selecting this training data, which could improve the detection performance in heterogeneous environments. In particular, improved classification results are first obtained with the proposed polarization-space 2-D Wishart classifier, which are then employed in a generalized inner product algorithm to select the secondary data. The proposed scheme is able to provide a better choice of secondary data, resulting in considerable improvement in detection performance. Numerical results are provided to show the effectiveness of the proposed method.