Abstract:This paper studied detecting soluble solid content （SSC）and pH of Bingtang orange by using near in－ frared diffuse reflectance. A total 45 samples were used to collect near infrared diffuse reflectance spectroscopy between 350 nm-1 800 nm， then 9 point moving average filter， the first derivative and multi-variant scattering correction were used to preprocess the primitive spectrum of Bingtang orange， and then the partial least squares （PLS），the principal component regression （PCR） and the multi-linear regression （MLR） were used to build prediction models which were used to quantitatively analyze SSC content and pH of Bingtang orange. The PLS model with the first derivative data treatment was prior to the other two ways based on the comparative analysis.