Abstract: In recent years, the subsynchronous oscillation (SSO) has frequently occurred in the wind power system. For the analysis and mitigation, it is of great importance to achieve a wide area monitoring of the incident. To achieve this, this paper proposed a new method to identify SSO parameters based on PMU data provided by wide area measurement systems. Through detailed mathematical derivation, it was revealed that the PMU data under SSO was dominated by four modes. The problem thus could be transformed to a modal parameter extraction (MPE) problem. Two widely-used MPE algorithms, matrix pencil method and the eigenvalue system realization algorithm were employed in this paper to identify the SSR parameters. In addition, the singular value decomposition and the coefficient of determination were used to improve the accuracy of the identification. The effectiveness and superiority of the proposed method was verified by synthetic signals, electromagnetic transient simulation and actual oscillation data of Guyuan, Hebei. The results indicate the proposed method can achieve good accuracy even at the initial stage of the incident; thus, it is a promising tool for the early warning of SSO.