The performance of a membrane deformable mirror with 37 electrodes (OKO Technologies) is studied in order to characterize its utility as an adaptive optics element. The control procedure is based on knowledge of the membrane’s response under the action of each isolate electrode, i.e., the influence functions. The analysis of the mathematical techniques to obtain the control matrix gives useful information about the surfaces that are within the device’s range of production, thus predicting the best performance of the mirror. We used a straightforward iterative algorithm to control the deformable membrane that permits the induction of surfaces in approximately four iterations, with an acceptable level of stability. The mirror and the control procedure are tested by means of generating Zernike polynomials and other surfaces. The mirror was incorporated in an adaptive optics prototype to compensate the eye’s aberration in real time and in a closed loop. Double-pass retinal images with and without aberration correction were directly recorded in a real eye in order to evaluate the actual performance of the adaptive optics prototype.