Abstract: Corn fiber is a kind of renewable biomass resources. In order to reliably analyze cellulose and hemicellulose of corn fiber, this work described the use of NIR spectroscopy to determined cellulose and hemicellulose content in corn fiber by PLS, the NIRS models for determination of cellulose and hemicellulose of corn fiber were set up by 1st derivative and second derivative smoothing data pretreatment methods. The results showed that for cellulose, the correlation coefficient(R) of calibration set(C-set) and validation set (V-set) were 0.9806 and 0.9799, standard error of calibration set(SEE) was 0.296323, standard error of cross-validation (SEP) was0.307204; and for hemicellulose, the correlation coefficient(R) of calibration set(C-set) and validation set (V-set) were0.9732 and 0.9005, standard error of calibration set(SEE) was 0.773057, standard error of cross-validation (SEP) was0.798132. The method could measure cellulose and hemicellulose in corn fiber rapidly and accurately, this method may provide a theoretical basis for the sustainable utilization of corn fiber.