With the development of hyperspectral remote sensing technology,hyperspectral prediction of crop growth can provide scientific management for agricultural production,which can improve crop yields and quality while avoiding excessive use of nitrogenous fertilizers.A mathematical model to invert the content of chlorophyll in cotton leaves was developed using continuous wavelet analysis and conventional spectral transformation to decompose and transform the raw leaf spectra of cotton.The characteristic wavelet coefficients and spectral characteristic bands were used as independent variables.Methods including univariate,stepwise regression and partial least squares were used.The results showed that different spectral treatments improved the correlation between the content of chlorophyll and spectral reflectance of cotton leaves.For the conventional spectral transformation,the inverse logarithmic first order differential lg(1/R′) improved the chlorophyll correlation of cotton leaves by 0.41.It is indicated that the continuous wavelet analysis is superior to traditional spectral models in terms of information noise reduction and mining of feature information.The model established has good stability with RPD>2 and good prediction ability for data sampled.