Polarized reflectance is the intrinsic properties of reflected light both by land surface and atmospheric particles. Therefore, polarized signals could be used to detect both surface and atmospheric information. Present scientific community mainly assumed that major information is contributed by atmospheric particles, regarding the surface contribution to be little. However, the sensitive research about how much the surface polarized reflectance could affect the TOA signals is rare. In this paper, we used the newly released POLDER datasets as the real surface polarized reflectance, the BPDF modeled polarized reflectance as comparison. The Monte Carlo simulations with huge number of iterations are carried out to compare the differences between the two datasets. The results suggest that the BPDF model could simulate better results for forest and bare lands, but relatively low accurate results for artificial lands. Our results indicate that surface polarized effects could not be neglected in polarized remote sensing.