Phycocyanin and chlorophyll-a are characterizations of cyanobacterial biomass.These two pigments are detected in situ with fluorescence spectrum method or fluorescence intensity.However,due to the fluorescence interference between phycocyanin and chlorophyll-a,the accuracy of in-situ detection of cyanobacteria is relatively low,which limited the application of this method.In this study,we developed a novel method of fluorescence analysis based on a two-order matrix model and a multivariate calibra-tion linear model for the correlation between the fluorescence intensity and pigment concentration,which was verified by experi-ment.The established method decoupled the two mixed pigments as well as quantitatively analyzed the wavelength selection of exci-tation and emission and largely eliminated the fluorescence interference between phycocyanin and chlorophyll-a,and therefore,sig-nificantly improved the accuracy of in-situ detection of cyanobacteria.More importantly,our study can provide theoretical basis for the cyanobacteria detection sensor in how to select wavelengths.