In the present study,the authors measured samples of typical forest soils in different states with multi-angle hyperspectral polarized reflections.The authors analyzed multi-angle hyperspectral polarized reflections of soil data with various viewing zenith angles,incidence angles,relative azimuth angles,polarized states,soil water content and soil granule.The authors found that those factors affected the reflectance values of forest soils but not the spectral feature.The conclusions included that the larger the incidence angles and viewing zenith angles are,the bigger the polarized reflectance values of the surface of the forest soil.When the forest soil was dry,the surface had phenomenon of diffuse reflection and the polarized light reflection did not take place.When the soil moisture content reached a certain level,the polarized reflection appeared.The more the moisture content of the forest soil was,the smaller the polarized reflectance of the surface.The bigger the soil granule was and the rougher the soil surface was,the smaller the surface polarized reflectance.The results and conclusions suggested that the spectral characteristics of the ground target need to be considered adequately in order to design the best mode for sensor systems by remote sensing technology.The authors suggest that the incidence angle and viewing zenith angle be selected on the basis of factual instance.The authors suggest using larger viewing zenith angles and that the incidence angle should be equal to the viewing zenith angle.In the meantime,the effects of sheltering by ground targets need to be considered and the proper state of polarization should be chosen while keeping relative zenith angle at 180°.This study not only helps find a new way for detection of soil characters,but also provides a theoretical basis for further research on multi-angle hyperspectral polarized reflection for detecting characteristic spectrum and best states in measuring forest soil.