In this paper, speed identification and control problems are simultaneously considered for a micro-grid power generation system (MG-PGS). To eliminate mechanical sensors, an observer has been designed to identify the speed of a wind energy conversion system based on a permanent magnet synchronous machine (PMSM). The proposed observer is an adaptive interconnected structure with online parametric estimation. Moreover, a new control based on the Lyapunov theory has been developed in order to improve sensorless control of a flywheel energy storage system based on induction machine. In fact, the main idea of this control technique is based on an appropriate choice of voltage control values that guaranty the global stability of the controlled system. Simulation results indicate that the sensorless control applied to the MG-PGS improves performances of the global control. The PMSM mechanical speed, the stator resistance, and inductance could be estimated precisely, and the dynamical property of the system is obviously improved.