This work explorers the feasibility of using hyperspectral technology to estimate the chlorophyll content of long-staple cotton leaves. The spectral data and chlorophyll content of long-staple cotton were measured from field observations; original spectral data were transformed to first-order differential and continuum-removed spectra, and then four common vegetation indices and three optimized spectral indices were calculated. According to correlation analysis, we used the support vector machines (SVM) regression model to estimate the chlorophyll content of long-staple cotton leaves. The results show that the chlorophyll content is negatively correlated with spectral reflectance in the ranges of 530?560 nm and 700?750 nm, and the correlation of the first-order differential spectra passed the 0.01 significance level test. Based on the first-order differential, an SVM model was established to estimate the chlorophyll content of long-staple cotton. The coefficient of determination reached 0.72, the root mean squared error and relative error were 1.99 and 0.72, respectively. This study demonstrates that the prediction accuracy is higher than SVM model constructed by the vegetation index and optimized spectral index. The results of this study can provide a reference for the rapid detection of chlorophyll content in long-staple cotton.