In view of the issue that the single-signal symbol rate estimation methods are unsuitable for time–frequency overlapped signals in underlay cognitive radio network (CRN), a novel symbol rate estimation method for time–frequency overlapped multiple phase shift keying (MPSK) signals in underlay CRN is proposed under low signal-to-noise ratio (SNR) conditions. In this method, the magnitude spectrum of teager energy operator (TEO) for the time–frequency overlapped MPSK signals is estimated first, and then, the nonlinear filter is employed to improve the spectral resolution of the TEO magnitude spectrum. Finally, the symbol rates of component MPSK signals are estimated by extracting the position information of the spectrum line, respectively. The modified Cramer-Rao bound of symbol rate estimation for the time–frequency overlapped MPSK signals is also derived in this paper. Simulation results are provided to confirm that the proposed method could estimate the symbol rate of the time–frequency overlapped MPSK signals effectively in low SNR regions.