The possibility of new technology occurrence has an important impact on the maintenance/ replacement decision. Therefore a challenge in maintenance decision marking is to determine maintenance policy and replacement investment plan under an uncertainty of technology improvement. In each period, we must decide whether to gather additional information on the potential improvement of a new technology, then chose the appropriate action for the asset (Do Nothing, Maintenance or Replacement). To formulate this scenario, we use a non stationary Markov Decision Process (MDP) model and provide some properties of the optimal policy based on a given set of numerical examples.