The proposed methodology is based on the developments of five significant points. These developments were necessary in order to give a balanced diagnosis. They represent however only one stage towards a better comprehension of the arising problem. The methodology is thus imagined to evolve with future scientific progress. These developments were made in the following fields : the positioning of the debris flow phenomenon among the other mass movements makes it possible to define it, in a univocal way, like "a lubricated granular flow, kinematically monophasic on a macroscopic scale" (chap. 2); the possibilities of connection between the pluviometry and the triggerring of debris flows were evaluated. This chapter analyses the difficulties of measurements of the rain in a mountainous area. These difficulties make possible to fix the degree of confidence to be given to the presented models (chap. 3); intuitively, there is a link between the debris flows and their catchment area. The relatively easy determination of the in situ characteristics (topography, geomorphology, geology) makes of this link a useful tool to estimate the danger. The presented works show the importance of surface, slope, geology, etc. to forecast which is the dominating phenomenon, volume, etc. (chap. 4); the release of the debris flows is primarily due to two mechanisms (erosion and fluidification of a sliding mass), which can appear according to several modalities. The knowledge of these mechanisms makes possible to delimit potential zones of release, on which the investigations will be concentrated (chap 5); to reflect field's observations, it is necessary to distinguish three families of debris flows. Each one has a proper constitutive's law. It is thus significant to distinguish between these families to be able to choose the appropriate propagation's model. Tools based on rheometrical tests are given to allow this choice (chap. 6). On the basis of this knowledge's body, it is possible to give a methodology in order to make a diagnosis of an alpine catchment area. This diagnosis is broken down into four stages. First of all the preliminary diagnosis aims to provide the minimum elements necessary for a coarse estimate of the danger. The three other stages can be made more or less in parallel and constitute the diagnosis itself. This diagnosis does not deal with propagation, but has to provide all the elements necessary to the models which will be used thereafter. Figure 7-5 summarizes this diagnosis.