With the increasing production and trade of surimi and surimibased foods,consumers have a higher quality requirement.Traditional analysis methods are timeconsumed,consumed large quantities of chemical reagents.Quality of surimi can be rapidly detected using nearinfrared analysis.Seawater surimi(including Alaska pollock surimi,hairtail surimi,sea bream surimi and mix surimi)and freshwater surimi(including silver carp surimi and grass carp surimi)were used in the study.Chemometric methods,including cluster and linear discriminant analysis(LDA)as well as partial least square(PLS)regression,were used to interpret spectral data.Results indicated that NIR method could successfully classify seawater and freshwater surimi with 100% prediction rate.In addition,the PLS models for water and protein content in surimi had good predictability:the correlation coefficients of the models were 0.98 for water and 0.96 for protein.Results showed that NIRS has great potential to be used in determining surimi quality.And important efforts for the practical application of nearinfrared have been made.In future studies,more representative samples should be added to enhance adaptability of models.