Most historians and philosophers of philosophy and history of mathematics hold one interpretation or the other of the nature of the method of analysis and synthesis in itself and in its historical development. In this paper, I am trying to prove – through three points – that, in fact, there were two understandings of that method in Greek Mathematics and philosophy, and which were reflected in Arabic mathematical science and philosophy; this reflection is considered as proof also of this double nature of that method. Thus, we have to rethink the nature of Arabic philosophy systems.