Efficient bioprocess control requires real-time monitoring of biomass, substrates, intermediates and nutrients. Presently used analytical methods, such as HPLC are time-consuming and require the use of environmentally unacceptable solvents. Unfortunately, the turbid nature of the fermentation broths inhibits direct use of traditional spectroscopic techniques because light scattering by the cells obscure the spectral features related to composition. In contrast to other spectral regions, the absorptions that occur in the near-infrared (NIR) region are weak and can be used to investigate samples that are highly absorbing and strongly light scattering. To demonstrate the applicability of NIR spectroscopy for bioprocess monitoring, we describe the spectroscopic determination of nutrient levels in fermentation broths by using diffuse reflectance measurements in the 1100-2500 nm region. A multiple linear least-squares regression equation was derived that compensates for scattering differences due to changes in the biomass of the fermentation broths. The spectroscopic model is as accurate as the reference analytical method used to derive the algorithm. The spectroscopic and chemometric principles used in developing the calibration models are discussed in detail.