Abstract : Multicriteria optimization consists in choosing, in the presence of multiple criteria, one (many) alternative(s) among an infinite number of alternatives which generally vary in a continuous domain. Since about thirty years, the field of multicriteria optimization knows a significant evolution. This evolution resulted in the development of a great number of methods. These multicriteria optimization methods are perceived like a richness of this field. Moreover, some justify it by the diversity of the problems and by the existence of various possible and legitimate resolution approaches of these problems. However, this phenomenon reveals also some weaknesses. Indeed, the majority of these methods miss axiomatic foundations, and it is difficult to choose the method to be applied to a given situation. In this thesis, we propose an axiomatic approach for multicriteria optimization. This approach is based on concepts such as the partial efficiency which are justified by understandable interpretations. It is robust with respect to the parameters and operational. It can be used in the resolution of various multicriteria situations such as the problems involving many incommensurable criteria and the public decision problems.