In this paper, we present a heuristic method to simplify the liquid crystal adaptive optics system (LCAOS) into a single-input-single-output (SISO) system, then build the dynamic model of LCAOS based on subspace identification. Results show that the identified model could accurately describe the dynamical behavior of LCAOS (97% match), with extremely low complexity. The wonderful features of low complexity and high precision, make the identified model highly beneficial for model based controller design, system analysis and dynamical behavior simulation of liquid crystal adaptive optics systems.