The great interest shown today in the double-fed induction generator in wind chain prompted us to analyze the performance of the separate control of active and reactive power generated by the stator of double-fed induction generator connected to the grid. The objective of the separate control of active and reactive stator powers is to provide constant power to the power grid whatever the variation of the wind speed, and to provide reactive power to correct the power factor. In this article, we present the modeling and an integrated vector control in a double-fed induction generator wind chain, which aims to independently control the active and reactive power generated by the double-fed induction generator. Knowing that the direct control of active and reactive power delivered to the grid with conventional proportional integral controllers provides good performance in permanent diet but a variable speed wind, the measured powers hardly follow the set values. To remedy this problem, we replace the conventional proportional integral controllers with proportional integral–fuzzy controllers to ensure good performance and robustness of the control of separate powers. To justify this choice, it is necessary to evaluate and compare the separate control loop for active and reactive power with conventional proportional integral controllers than with proportional integral–fuzzy controllers under the same constraint variable speed wind. The second part treated in this article is to reduce the harmonic pollution caused by the supply of the double-fed induction generator–based inverters and controlled rectifiers, which strongly infect current line. This phenomenon can have an adverse effect on the loads connected to the grid and equipment of proximity. The insertion of a filter system in the chain is needed, except that the strain of the volume and the reflection require a high cost for a more advanced and more optimal solution. Our contribution to this part is to replace the supply of the double-fed induction generator rotor side by a matrix converter controlled by the modulation technique of the difference errors between the measured values and the desired output values (least mean square errors).