Abstract: Abstract: The aim of this study was to investigate the effects of freeze-thaw cycles on the quality of large yellow croaker by physico-chemical method and evaluate a rapid method that was based on front-face fluorescence spectroscopy (FFFS) and used to differentiate between fresh and frozen-thawed samples. The pH value, thawing loss, color parameters, thiobarbituric acid reactive substances (TBARS), and carbonyl content of large yellow croaker muscle were measured after the treatment of different freeze-thaw cycles (0, 1, 3, 5 and 7 times). The results showed the thawing loss of the fish flesh increased significantly (P<0.05) and the pH value showed a trend of first increasing and then decreasing (P<0.05) as the number of freeze-thaw cycles increased. Freeze-thaw cycles provoked a decrease (P<0.05) in CIE a*, and an increase (P<0.05) in CIE L* and CIE b*. According to the TBARS values and carbonyl content determined, lipid and protein were oxidized significantly (P<0.05). The above results indicated that the repeated freeze-thaw cycles during storage had a detrimental effect on the qualities of large yellow croaker. The fluorescence emission spectra of tryptophan residues (excitation: 290 nm; emission: 305-400 nm) of proteins and nicotinamide adenine dinucleotide (NADH) (excitation: 336 nm; emission: 360-600 nm) were adopted on treated samples. The tryptophan residues fluorescence emission spectra recorded following excitation at 290 nm showed similar shapes among samples with a maximum at about 330 nm. The NADH emission spectra (excitation 340 nm) of all fish samples showed a difference at 465 nm between fresh and frozen-thawed samples. Principle component analysis (PCA) method was applied to tryptophan residues and NADH fluorescence spectral data sets. The first 2 factors of tryptophan residues fluorescence spectra explained 97% of the total variance and the first 3 factors of NADH fluorescence spectra explained 93% of the total variance. PCA results showed a good discrimination between fresh and frozen-thawed fish samples. However, fluorescence spectroscopy combined with PCA could not discriminate clearly between fish samples which had been treated under different freeze-thaw cycles. Fisher linear discriminant analysis (FLDA) method was performed on the 2 data sets. The first 2 functions of tryptophan residues and NADH fluorescence spectra allowed to explain 98.5% (function1 95.7% and function2 2.8%) and 90% (function1 55.5% and function2 34.5%) of variance, respectively. From the NADH spectra, FLDA gave a better distinction between fresh and frozen-thawed fish samples than PCA. Separation between the frozen-thawed samples depending on freezing cycles was observed from the similarity map determined by discriminant factors. Discriminant model of fresh and frozen-thawed large yellow croaker was set up through FLDA. Considering tryptophan fluorescence spectra, correct distinction of 68.3% and 66.7% was observed respectively for the original and cross validation spectra. A better classification was obtained from NADH fluorescence spectra since 100% correct classifications were obtained for the original and cross validation spectra. It is concluded that NADH fluorescence spectra may be considered as a promising method for the reliable differentiation between frozen-thawed and fresh fish. Our results show that front-face fluorescence spectroscopy coupled with chemometric methods has great potential in the development of rapid and non-destructive methods for the freshness evaluation of aquatic products.