Dynamic planning concerns the planning and execution of actions in a dynamic, real world environment. Its goal is to take into account changes generated by unpredicted events occurred during the execution of actions. In this paper we develop the theoretic model of dynamic planning presented in [12]. This model proposes a graph representation of possible, efficient and best plans of agents acting in a dynamic environment. Agents have preferences among consequences of their possible actions performed to reach a fixed goal. Environmental changes and their consequences are taken into account by several approaches proposed in the so-called "reactive planning" field. The dynamic planning approach we propose, handles in addition changes on agents´ preferences and on their methods to evaluate them; it is modeled as a multi-objective dynamic programming problem.