Aiming at virtual air data estimation without initial wind velocity, a two-step air data estimation method aided by the aerodynamic model and navigation information is proposed. On the foundation of dynamic equations with the aerodynamic model, the measuring model of navigation sensors directly relate to wind speed is established. The nonlinear least squares optimization algorithm is used to estimate the wind speed roughly, and take the estimation as initial value. The precise estimate of wind speed is achieved with the extended Kalman filter(EKF). Based on the wind estimation, the real-time and high accuracy estimation of air data is realized. The simulation results show that, this algorithm has higher convergence speed and estimation accuracy, and improves the air data system's measurement range and reliability, which can apply to the measuring of true air speed, angle of attack and sideslip within full flight envelope.