Based on policy change literature and in an improved version of D. Beetham's (1991) model for legitimacy (Alagappa, 1994), I propose here an approach for the study of policy stability & change (called legitimation & policy dynamics). Oriented to explain policy change in political systems defined by its institutional fragility and persistent legitimacy deficits, LPD is an actor-centered perspective, in which legitimation of power through policy is assumed as an unavoidable task, and conforms as a causal-driver useful in explaining policy stability and change. LPD assumes that policy change can take two forms: as a reactive way or as a proactive logic. In both of these forms the actors of the dominant coalition will seek to maintain an active presence (and increase own’s influence and control capabilities if possible) over policy and its change processes. Institutionally conditioned, these actors may assume four differentiated operational positions (shock response, strategic improvement-based, thermostatic and change-contention) and in doing so they also configure narratives and send clear messages that influence all actor expectations during change process. The “legitimacy pattern” associated with a given policy design favors periods of stability based on perpetuation, the logic of adaptation and incremental changes or planned and long-time based processes of policy change. Focusing events, external shocks, innovation and diffusion processes, or endogenous dynamics are all forces that influence the policy subsystem and may lead to distortions (based on unconformity with rules, discrepancy with shared beliefs, withdrawal of consent or an inadequate policy performance) in the legitimacy pattern. Those distortions may force changes in the dominant coalition and in public policy, but it is only with a transformation of the legitimacy pattern when a major policy change occurs. Preliminary hypotheses are here proposed.