When the observed input-output data are corrupted by observed noise in an aircraft flutter statistic
model, one should obtain accurate parameters of the model. We combine the instrumental variable identification
method and covariance matching method to develop a new method: the instrumental variable covariance
method. In the statistic model of aircraft flutter, we introduce some instrumental variables to develop a covariance
function. We present the procedure to solve the criterion function and partial derivatives expression. We
derive the asymptotic covariance matrix expression obtained with the proposed method, and use the asymptotic
covariance matrix expression to verify effectiveness of the method, and design the external excitation
signal. The method is applied to identify the transfer function in the current loop of a flight simulator and
parameters of an aircraft flutter model. The simulation shows effectiveness of the method.