Near-infrared (NIR) prediction model of ramie pectin content was established in this research. Wet chemical analysis method which was based on Chinese national standard was conducted for getting calibration data, and NIR data of 60 ramie samples were collected using acousto-optic tunable filter (AOTF) near-infrared spectrometer. The NIR model of pectin of ramie were derived by partial least square (PLS) regression. Prediction of chemical composition of independent ramie samples showed that R/SEP values of pectin is 17.87. Such NIR calibration model can be utilized by ramie fiber manufacturers and breeding workers, in order to better manage the degumming process and evaluate the quality of ramie varieties.