This study reports the first combined use of a fluorescence analytical spectroscopy, guest-host chemistry and multivariate regression analysis for rapid simultaneous determinations of cysteine and homocysteine thiolactone (homocysteineTL) (potential markers for various diseases including the Alzheimer's disease and cardiovascular diseases) in human serum albumin (HSA) at physiologically relevant conditions. The binding of homocysteineTL with HSA in the phosphate buffer, pH 7.4 or tris-borate buffer, pH 10.0, results in a dramatic quenching of the fluorescence emission property of HSA. In contrast, the binding of cysteine with HSA has no significant influence on the emission property of HSA, suggesting differences in the binding mechanism of cysteine and homocysteineTL with HAS. The results of the FTIR experiments indicate that hydrogen bonding and disulfide covalent bonding play a considerable role in cysteine-HSA and homocysteineTL-HSA complexation. A partial-least-square (PLS-1) regression multivariate analysis was employed to correlate changes in the fluorescence spectra of solutions with varying concentrations of cysteine and homocysteineTL mixture in HSA samples. The PLS-1 regression models were later used for simultaneous determinations of cysteine and homocysteineTL concentrations in HSA validation test samples. The developed PLS-1 regression models were able to predict the concentrations of cysteine and homocysteineTL in HSA validated test samples with reasonable good accuracy and a low root mean square percent-relative-error-of-prediction of 4.11% for cysteine and 5.66% for homocysteineTL.