Moisture content is an important trait for rubber sheet trading system. Therefore, a calibration equation for predicting moisture content was created by near infrared (NIR) technique in order to develop a more fair trading system in Thailand. Spectra were recorded in two systems. One was measurement on each rubber sheet and the other was on a pile of sheets. Both were measured by a handheld NIR spectrometer in the short wavelength region (700–1100 nm) in the transflectance mode using Teflon as a diffuse reflector. The spectra showed the peak at about 900 nm which belongs to isoprene, the major component of rubber sheet. Pretreatment with second derivative was applied to remove baseline shift effect occurring due to thickness differences on each rubber sheet. From validation results, moisture contents predicted by single sheet system were more accurate than a pile of sheet system with standard error of prediction (SEP)=0.39% and bias of -0.07%, and they were not significantly different from the actual values at 95% confidence. As a result, determining moisture content in each rubber sheet by a handheld NIR spectrometer provided accurate values, easy and rapid operation.