The study aims to rapidly estimate chlorophyll content in plants by polarized light remote sensing and apply the method in agrometeorological automatic observation. According to the physical characteristic that the polarized response is produced by chlorophyll particles, SPAD-502 chlorophyll meter and FieldSpec^®3 Hi-Res portable hyper-spectrometer with polarizer were used on the multi-angle observation platform to measure and study the relationship between polarized hyperspectral characteristics and the chlorophyll content of three representative plant leaves with smooth surface, including mulberry, photinia frase and camellia. The results indicate that the intensity of green peak of smooth leaves near 550nm decreases gradually with the increase of chlorophyll content. In other words, the polarization hyperspectrum has a non-linear relationship with chlorophyll content. The Degree of Polarization(DOP) of green peak band has the best relationship with chlorophyll content, followed by polarization reflection, with the highest reflection and total reflection behind, and the relationship with the lowest reflection is the least. This study has great significance for further study on evaluation of crop growth status, disaster and yield estimation with polarized remote sensing.