Energy efficiency is an important issue. Innovative technologies are developed in order to make buildings better at energy efficiency. Energy efficiency labels are used around the world to show how much energy efficient a building is. This labelling process is an assignment of buildings into predefined classes. If we consider the assignment problem has multiple criteria to consider and the classes are ordered, this problem can be handled by using multiple criteria sorting techniques. Furthermore, Inverse Multiple Criteria Sorting Problem (IMSCP) is concerned with the selection among the possible actions, which can change objects state in terms of criteria to obtain a better sorting of objects. In this study, a linear programming model of IMSCP is proposed to construct an energy efficiency improvement plan for buildings. The main aim is to choose the actions, which give us the desired labels at minimum cost. An illustrative example is presented to demonstrate the applicability of the proposed model. Solution results prove that this model is appropriate for energy efficiency improvement of buildings.