The application of neural network based schemes for parameter estimation of unstable/augmented aerospace dynamic systems is considered. The parameter estimation algorithms are programmed within the structure of recurrent neural networks (RRNs). Numerical simulation results are given.