The estimation of crop chlorophyll content could provide technical support for precision agriculture. Canopy spectral reflectance was simulated for different chlorophyll levels using radiative transfer models. Then with multiperiod measured hyperspectral data and corresponding chlorophyll content, after extracting six wavelet energy coefficients from the responded bands, an evaluation of soybean chlorophyll content retrieval methods was conducted using multiple linear regression, BP neural network, RBF neural network and PLS method. The estimate effects of the three methods were compared afterwards. The result showed that the three methods based on wavelet analysis have an ideal effect on the chlorophyll content estimation. R2 of validated model of multiple linear regression, BP neural network, RBF neural network and PLS method were 0. 634, 0. 715, 0. 873 and 0.776, respectively. PLS based on Gaussian kernel function and RBF NN methods were better with higher precision, which could estimate chlorophyll content stably.