A soft-sensor for monitoring solubility of native-like a-lactalbumin (alpha-LA) and beta-lactoglobutin (beta-LG) and their aggregation behavior following heat treatment of mixtures under different treatment conditions was developed using fluorescence spectroscopy data regressed with a multivariate Partial Least Squares (PLS) regression algorithm. PLS regression was used to correlate the concentrations of alpha-LA and beta-LG to the fluorescence spectra obtained for their mixtures. Data for the calibration and validation of the soft sensor was derived from fluorescence spectra. The process of thermal induced aggregation of beta-LG and a-LA protein in mixtures, which involves the disappearance of native-like proteins, was studied under various treatment conditions including different temperatures, pH, total initial protein concentration and proportions of alpha-LA and beta-LG. It was demonstrated that the multivariate regression models used could effectively deconvolute multi-wavelength fluorescence spectra collected under a variety of process conditions and provide a fairly accurate quantification of respective native-Re proteins despite the significant overlapping between their emission profiles. It was also demonstrated that a PLS model can be used as a black-box prediction tool for estimating protein aggregation when combined with simple mass balances.