A current trend in modern near-infrared spectroscopy is the incorporation of sophisticated mathematical algorithms into the computer instrumentation used to extract information from raw spectral data by applying complex multivariate models. To address some of the problems that near-infrared spectroscopy faces, the GrainNet software model that connects a MATLABRTM computing and development environment, NIR spectrometers, and MS Server data-storage for spectral data and calibration models, was developed.;GrainNet is a client-server based Internet enabled communication and analyzing model for Near-Infrared (NIR) instruments. FOSS Infratec, Perten, and Bruins Instruments are currently three brands of the NIR instruments that have been included in the project. The performance of the implemented calibration models was evaluated. Three calibration models are implemented in the GrainNet: (1) Partial Least Squares Regression; (2) Artificial Neural Network; (3) Locally Weighted Regression.;The Piecewise Direct Standardization (PDS), Direct Standardization (DS), Finite Impulse Response (FIR) and Multiplicative Scatter Corrections (MSC) models were developed in the MATLABRTM environment and tested for standardization transfer of the Bruins Instruments and Foss Infratec grain analyzers. A new calibration model for corn that uses feed-forward back-propagation neural networks with wavelets signal decomposition used as an input was developed.