Red tides can form as a result of explosive increases in the biomass of some algae in marine or freshwater environments, posing a threat to ecological security and socio-economic development. During a red tide outbreak, most of the algae in the water column are present in a mixed and coexisting form of multiple species. In practice, fluorescence between different algal species not only overlaps linearly, but also quenches due to mutual absorption, making multi-component identification and quantitative analysis of algae difficult to achieve an accurate level. In this paper, we take the common red tide algae species in natural water bodies as an example, and take different concentrations of pure and mixed algal samples to measure their excitation fluorescence spectra at the characteristic excitation wavelength of chlorophyll a. Then we apply the Crosstalk Fluorescence Spectroscopy Analysis (CFSA) algorithm to the identification and quantitative analysis of mixed red tide algae excitation fluorescence spectra, and the quantitative analysis accuracy of this algorithm (R² > 0.93) is higher than that of the traditional multiple linear regression algorithm. The algorithm confirmed the feasibility of using the excitation fluorescence spectrum and the CFSA algorithm to classify and measure the concentration of red tide algal species. The accurate identification and quantitative analysis of mixed red tide algae can be used for early warning of water quality, promoting the construction of a harmful algal bloom monitoring and warning system, and better addressing the problem of harmful algal blooms.