Abstract: Subsynchronous oscillation in power systems with high penetration of renewables and inverters is mainly caused by the resonance of the power electronic equipment of wind power and photovoltaic, which may be accompanied by frequency coupling of supersynchronous oscillation. Oscillations spread widely and change quickly. Therefore, synchronous dynamic monitoring of subsynchronous oscillation is required. In this paper, a dynamic parameter identification method of power system sub/supersynchronous oscillations based on synchronous phasor spectrum fitting is proposed. By constructing a system of equations in matrix form, the superposition characteristics of the synchronous phasor spectrum are restored. It can accurately obtain the frequency, amplitude and phase of the basic component of the frequency shift and the sub/supersynchronous components. This algorithm has great advantages over the existing algorithms. First, the algorithm significantly reduces the necessary time window for spectrum analysis methods to 200 ms. It can achieve dynamic synchronous monitoring of sub synchronous oscillations in the order of one hundred milliseconds using fundamental synchronous phasors. Then, the problem of the influence of synchronous phasor data loss on parameter identification is solved. Finally, this algorithm uses spectral amplitude as error weight to improve the accuracy of parameter identification results. The results of simulated synchronous phasor data and actual simulation data show that the proposed method is still effective even under the condition of noise and data loss.