Abstract: With the rapid advancements in AC-DC hybrid transmission technology and the growing prevalence of power electronic equipment in power systems, there is an escalating need to analyze and study the electromagnetic transient simulation of these systems. However, modeling the electromagnetic transients in actual large power systems proves to be both complex and time-consuming. To address this issue, a novel approach based on combined deterministic-stochastic identification is proposed for external system transient equivalence in power systems. Initially, a deterministic excitation system response model is established using input and output data from the external system. This model is developed via a deterministic random subspace method, allowing for accurate resolution of the state-space expression of the system. Subsequently, this state-space expression for the external system is converted into an admittance rational function expression. An RLCG equivalent circuit is then constructed based on the zero-pole locations, simplifying the modeling process and enabling the construction of a transient equivalent for the external system across a wide frequency band. In the final stage, simulation and measured data are employed to validate the proposed method. The findings demonstrate that this new approach can significantly reduce simulation time and accelerate the overall simulation process.