The aim of multi-objective optimization or multiple-attribute decision making is to find the most appropriate solution among Pareto solutions. Usually, decision makers are in charge of evaluating and rating objective functions or attributes. However, in a good number of practical decision making problems especially in the area of public services e.g. determining the service contents of public transport, the decision makers determine the service contents while the users evaluates the quality of the service. In these problems where decision makers and evaluators are different, an effective method is necessary that can estimate the preferences of the evaluators (users). In the paper, an interactive method is proposed that estimates or learns the user preferences and finds the solution based on the estimated results.