The present paper firstly denoises the signal with morphological method, selecting sine-shaped structure element, using the morphological difference in waveform between the three-dimensional fluorescence and noise signal, then singular value decomposition is applied to the denoised data, and finally the chemical rank is determinated jointing eigenvalues and eigenvectors form singular value decomposition. This paper principally discusses the theory basis of morphological filtering method, firstly simulated data is analysed by morphological filtering method to confirm the necessity and effectiveness of proposed method, then the feasibility and practicability of the proposed method is verified by the determination of components number of phenols mixture three-dimensional fluorescence spectra compared with traditional Monte Carlo method. The experiments demonstrate that the proposed method is able to estimate the chemical rank correctly.