Abstract: To improve the sharpness of polarimetric synthetic aperture radar (PolSAR) images and keep the polarimetric characteristics of targets after filtering, Lee filter of PolSAR image based on typical scattering difference index （TSDI） is proposed. According to the TSDI between pixel to be filtered and its neighboring pixels, an adaptive threshold method is used to select homogeneous pixels of the filtered pixel, which is used for Lee filter. For selecting threshold, the probability distribution of the TSDI is firstly estimated by Parzen window, based on which the threshold is calculated. The PolSAR data acquired by the AIRSAR and the UAVSAR system is used to illustrate the effectiveness of the proposed algorithm. Experimental results show that compared with the refined Lee filter, the proposed algorithm reduces the spackle noise more thoroughly, as well as keeps the edge sharpness of images and the polarimetric characteristics of targets better.