Additive manufacturing (AM) is emerging as an important manufacturing process and a key technology for enabling innovative product development. Design for additive manufacturing (DFAM) is nowadays a major challenge to exploit properly the potential of AM in product innovation and product manufacturing. However, in recent years, several DFAM methods have been developed with various design purposes. In this paper, we first present a state-of-the-art overview of the existing DFAM methods, then we introduce a classification of DFAM methods based on intermediate representations (IRs) and product's systemic level, and we make a comparison focused on the prospects for product innovation. Furthermore, we present an assembly based DFAM method using AM knowledge during the idea generation process in order to develop innovative architectures. A case study demonstrates the relevance of such approach. The main contribution of this paper is an early DFAM method consisting of four stages as follows: choice and development of (1) concepts, (2) working principles, (3) working structures, and (4) synthesis and conversion of the data in design features. This method will help designers to improve their design features, by taking into account the constraints of AM in the early stages.