The application of vector control to an induction motor depends largely on the precision with which used parameters are known. The variables and parameter estimation that are not directly measured lead us to construct a state observer. The presence of noise introduced by the inverter, model and sensors suggests the use of a filter. In this paper, we analyze the indirect method of the vector control. We propose the utilization of an Extended Kalman Filter (E.K.F) to build an adaptive version of control by oriented flux of double-star asynchronous machine (D.S.A.M). The effectiveness of the resulting filter is analyzed on the basis of several numerical simulations.