Spatially resolved spectroscopy (SRS), based on near-infrared, is better adapted to extract physical information (here, particle size) than classical spectroscopy in heterogeneous media (suspension, emulsion, powder, foam, etc.). It was employed to monitor the particle size in emulsion copolymerization. The prediction of the particle size was distinguishable from the prediction of the polymer content, based on their different impact on the spectra due to scattering and absorption phenomena. Models based on the partial least-squares regression were developed. Different types of mean diameters were studied to explore the sensitivity of the SRS and the possibility to reconstruct the particle size distribution. Raman spectroscopy was also investigated to determine the impact of particle size on the signal and how far it can be quantitative.