The rapid and nondestructive estimation of leaf chlorophyll content is significance of the monitoring of vegetation growth and environmental stress. The traditional method of chlorophyll estimation is mostly based on the spectral information of upper leaf side. However, in the actual remote sensing observation, the sensor not only receives the adaxial spectral information of leaves, but also receives the spectral information from the abaxial surface of leaf. The main purpose of this study is to find an accurate method to estimate the chlorophyll content of leaves when considering both the adaxial and abaxial spectral information of the blade. This paper compared the simple difference vegetation index (SD), simple ratio vegetation index (SR), normalized difference vegetation index (ND) and partial least squares (PLS) regression modeling method. It was found that PLS regression modeling method had the highest precision in all of the methods to estimate leaf chlorophyll content of two species with two surfaces reflectance. The R2 was 0.91 and the RMSE was 5.21 g·cm-2 Therefore, it can be concluded that the PLS method is more accurate in estimating leaf chlorophyll content when considering both adaxial and abaxial leaf spectral information.