In this paper I propose a method to determine portage routes using mainly archaeological and environmental data, through the example of the portage between two French rivers in ancient times. During the Roman period, one could reach the Rhine frontier through several routes from the city of Lyons. These routes connected various large drainage basins though portages. The one between the Saône and the Moselle rivers was particularly important. Yet, we still do not know its exact path. This method of investigation involves two aspects. First, I need to determine the main terminal-points of navigation on the two rivers. Due to the lack of data in ancient texts, the sources I am using are works on the hydrology of the Saône and Moselle rivers, archaeological remains and modern texts. This will allows me to propose several possible transloading points according to the environment in the period, to the seasons and to the cultural context. Second, I simulate routes between the proposed terminal-points, which enables me to distinguish two kinds of paths: on the one hand, there was a long portage, departing from Chalon-sur-Saône or more upstream; on the other hand, there was a short portage departing much more upstream on the Saône river. The short portage could only be taken when environmental conditions allowed, whereas the long one was passable all year long. The modeling of the path of least cost between the various terminal-points sheds new light on the various possible options and the choices the transporters made. It appears that choosing a short and cheap portage or a long and expensive one depended not only on the topographical, river and climate constraints but also on the objectives of the journey and the commercial strategies of the merchants.