The wide area measurement system（WAMS）and synchronous phasor measurement unit（PMU）have been widely deployed in new energy stations above 40 MW according to relevant requirements. Real-time data obtained through their high-precision sampling provides conditions for the dynamic modeling and identification of new energy units. Standard units provided in the PSD-BPA transient stability program user manual is used to establish a parameter electromechanical transient model for photovoltaic power generation systems.Based on prediction errors, a closed-loop identification method uses WAMS data to identify and analyze the parameters of the photovoltaic power generation system online. On this basis, a prediction error method combined with the box-jenkins（BJ）model is proposed to identify the uniqueness of parameters. A photovoltaic power generation system model is constructed through a simulation platform, and the measured WAMS data of the photovoltaic power station in Henan power grid is fitted and analyzed, fully verifying the effectiveness of the proposed method for closed-loop identification.