We describe a strategy for the locking of high-finesse suspended optical cavities taking advantage from the knowledge of mirrors dynamics, at least in an approximate way. We argue that the considered locking procedure can be more efficient than the naive one. We discuss possible generalizations and we point to future investigations.