Small fluctuations caused by random changes of loads exist continuously in power grids，which are referred to as ambient signals.Based on wide area measured ambient data，a method for a closed-loop identifying the power system model based on ambient data is proposed.This model can be used to accurately identify the current operating conditions of a power system.It also provides useful information for system analysis and controller design tasks.The feasibility of closed-loop identification of the system model based on the auto regressive moving averaging vector（ARMAV）model is discussed to show that the power system model can be identified from multiple ambient signals.The method is applied in a two-area four-machine system and a36-bus system.The results validate the correctness of the proposed method.