Adaptive Optics (AO) refers to (servo-control) methods used in astronomical imaging to compensate in real time for the loss of spatial resolution power of the instrument caused by the earth's upper atmosphere turbulence, which produces inhomogeneties of air refractive index and hence modifies in a chaotic manner the optical phase of incoming light from space. In this paper we present the first results of a radically new approach to phase reconstruction in AO based on non linear methods for the processing of complex signals having multiscale properties. We make use of the Microcanonical Multiscale Formalism (herein referred to as MMF) to infer properties along the scales of the complex signal made of the perturbated optical phase, and we perform the reconstruction of the phase using appropriate wavelet decomposition associated to the cascading properties of fully developped turbulence.