This study demonstrates the application of fluorescence excitation-emission matrix (EEM) spectroscopy to the quantitative predictive analysis of recombinant glycoprotein production cultured in a Chinese hamster ovary (CHO) cell fed-batch process. The method relies on the fact that EEM spectra of complex solutions are very sensitive to compositional change. As the cultivation progressed, changes in the emission properties of various key fluorophores (e. g., tyrosine, tryptophan, and the glycoprotein product) showed significant differences, and this was used to follow culture progress via multiple curve resolution alternating least squares (MCR-ALS). MCR-ALS clearly showed the increase in the unique dityrosine emission from the product glycoprotein as the process progressed, thus provided a qualitative tool for process monitoring. For the quantitative predictive modelling of process performance, the EEM data was first subjected to variable selection and then using the most informative variables, partial least-squares (PLS) regression was implemented for glycoprotein yield prediction. Accurate predictions with relative errors of between 2.3 and 4.6% were obtained for samples extracted from the 100 to 5000 L scale bioreactors. This study shows that the combination of EEM spectroscopy and chemometric methods of evaluation provides a convenient method for monitoring at-line or off-line the productivity of industrial fed-batch mammalian cell culture processes from the small to large scale. This method has applicability to the advancement of process consistency, early problem detection, and quality-by-design (QbD) practices.