Technical Abstract: When NIR spectroscopy is used as a quantitative analysis technique, spectra are collected for a calibration sample set, the chemical analysis are entered into the computer for this calibration set, and the calibration model or equation is computed. When these steps are completed, the calibration is tested for its accuracy in measuring the chemical composition for unknown samples using a set of validation or test samples. If the accuracy is not sufficient, the process is repeated by including a wider variation of sample types in the calibration, and by double-checking the accuracy of the laboratory used for the initial manual wet laboratory results. This process is repeated until the validation or test set yields accurate results. The purpose of the calibration model is to relate the concentration of some analyte found in a sample to the spectral data collected from that sample. The sample sets used to develop calibrations must comprise a wide range of analyte values as well as spectral responses. Instrumental artifacts are convolved into the near-infrared (NIR) spectra, and thus unless instruments have identical response across systems (or following major repair) the instrument itself must be considered as well when developing a calibration model. While it is true that some NIR instruments have calibrations for specific analytes built into the instrument itself, a similar process was performed at some time in order to develop the on-board calibration. This chapter is concerned mainly with the philosophy of the process of calibration development rather than emphasizing the chemometric methods available.