In order to explore a better classification strategy for multi-band polarimetric SAR, we apply both the deep learning classifier and the clustering method based on scattering mechanism to the airborne polarimetric SAR data of P, L, S, $\mathrm{C}$, X band. Firstly, the characteristics of deeplab $\mathrm{v}3+$ segmentation results in different frequency bands are analyzed. Then, the performance of the clustering results based on the h-alpha-wishart method in different frequency bands is analyzed. Through the comparison of classification results, we summarize the characteristics of different frequency bands, which provides support for subsequent classification strategy optimization.