In this thesis, a blind system identification technique using a Wiener model is used to estimate the power system modes. A Wiener model, a universal approximator consisting of a dynamic linear system followed by a memoryless nonlinear element, is used to estimate the power system nonlinearities. It also constructs a set of intermediate data, which can be used by a linear estimation technique such as subspace identification for estimating the power system electromechanical modes. In this research, a blind variant of a subspace method, Numerical Algorithm for Subspace State Space System IDentification (N4SID), is used. The algorithm is tested with simulation data from the Kundur two-area network. The accuracy and reliability of these estimates are accessed by carrying out Monte Carlo simulations. The estimated results obtained from the simulated system using a Wiener model were very accurate with reduced prediction errors and was a good fit for the power system.