Chlorophyll content in leaves is one of the important internal information for predicting plants growth status. In this study, we use near infrared (NIR) spectroscopy technique to predict chlorophyll content in pepper leaves. Calibration models were created from spectral and constituent measurements, chlorophyll content measured by a SPAD-502 chlorophyll meter, 74 samples served as the calibration sets and 16 samples served as the validation sets. Partial least squares (PLS) and principal component regression (PCR) analysis technique were used to develop the prediction models, and four different mathematical treatments were used in spectrums processing: smoothing, baseline correction, different wavelength range, first and second derivative. When we use PLS analysis and select spectra with second derivate, we can get high correlation efficient and low RMSEC value, but big difference between RMSEC and RMSEP. The best calibration model when we delete four outlier samples, when we process spectra with second derivate at full wavelength, we can get highest correlation coefficient (r=0.97537), a relative lower RMSEC value (2.33), and a small difference between RMSEC (2.33) and RMSEP (5.49). Result showed that NIR technique is a non-destructive way; it can acquire chlorophyll content in pepper leaves quickly and conveniently.