Complex clutters of High Frequency Surface Wave Radar cause difficulties on target detection. Based on polarization characteristics of HFSWR, a new method by combining polarization information with the traditional dynamic programming algorithm is proposed. Moving target direction information and polarization information are used to improve the performance of the algorithm, and pre-detection technique is adopted to reduce the amount of calculation. The simulation and test results in a Gaussian background and HFSWR measured data show that the new algorithm based on polarization information has significant performance benefits compared to the traditional method based on dynamic programming algorithm in the ionosphere clutter.