Abstract : We present a blind recursive algorithm for the estimation of nonstationary SIMO systems. An adaptive algorithm is developed to track the noise projector, to evaluate the degree of intersymbol interference (ISI), and the channel parameters. This method is also capable of handling abruptly changing channels during the observation interval. We define a new filtering operator that matches this projector with the subspace method. Finally, we relate this work with Capon's approach and the beamforming method. Simulation results concerning the performance of the developed method and its comparison with the classical subspace method are presented.