Abstract: Hyperspectral remote sensing reflectance data and suspended sediment concentration (SSC) data of the Pearl River on February and May 2014 was used in the paper to establish two band remote sensing retrieve models. By comparing of February remote sensing retrieve model with May2014remote sensing model, we found the SSC and spectral reflection existed low correlation(R²=0.422) in winter but had a strong correlation in spring(R²=0.716). May is a wet season of the Pearl River that the runoff increase and bring more sediments into river. At the same time fresh water increase and create upwellings so that caused sediment resuspending. Otherwise there exists brackish water front and salinity gradient near it. They also could lead to SSC increase. SSC is sensitive to light that high SSC value could run model well. On the contrary, February is a dry season of the Pearl River, the major SSC comes from runoff. The reduction of runoff leads to low SSC, which has bad effects on model.