The scope of this presentation is a new methodology to correct conventional NIR data for scattering effects. The technique aims at measuring the absorption coefficient of the samples rather than the total attenuation, measured by conventional NIR spectroscopy. The main advantage of this is that the absorption coefficient is independent of the path length of the light inside the sample, and therefore independent of the scattering effects. The measurements in this work were made using a novel system for time-resolved measurements, based on short light continuum pulses generated in an index-guided crystal fibre and a spectrometer-equipped streak camera. The system enables spectral recordings in the wavelength range 500 - 1200 nm with a spectral resolution of 5 nm and a temporal resolution of 30 ps. The evaluation scheme is based on modeling of light transport by diffusion theory, that provides an independent measure of the scattering properties of the samples, that later is used to correct conventional NIR data. This yields a clear advantage over other pre-processing techniques, where scattering effects are estimated and corrected for by using the shape of the measured spectrum only. PLS calibration models shows that, by using the proposed evaluation scheme, the predictive ability is improved by 50 % as compared to models based on conventional NIR data. The method also makes it possible to predict the concentration of active substance in samples with physical properties different from those of the samples included in the calibration model.