The magnetic fluid deformable mirror (MFDM) is a new type of wavefront corrector which has advantages compared to conventional wavefront correctors, such as higher stroke and lower fabrication cost. In this paper, a Youla (Q) parameterized adaptive regulation approach is proposed for the MFDM to correct both random and deterministic dynamic aberrations based on the dynamic model of the MFDM. The parameter Q is tuned online by a recursive least squares (RLS) algorithm so that it converges to the desired value. An adaptive optical system with a prototype MFDM is setup to verify the effectiveness of the adaptive controller. The control performance is also compared with a H∞ controller to show the improvement of the wavefront residue RMS error. The result illustrates that the proposed adaptive control method can effectively control the mirror surface for the unknown dynamic aberration correction, especially to the deterministic disturbances in adaptive optics systems.