Significant progress was made these last decades in the development of hydrogeological methods. However, these new methods are still very little used in the problematics of unstable slopes, mainly because of the hydrogeological complexity, which characterizes them (thin aquifers, discontinuous media, succession of saturated and unsaturated zones, lower permeability, higher hydraulic gradients, evolution of the permeability in time in relation to slope movements, etc.). Consequently, the approach usually developed to study slope instabilities is based on the setting-up of relations between hydroclimatologic parameters and slope movements. This approach exclusively concentrates on causes and effects, without developing the processes induced by these causes and responsible for these effects. This "black-box" type of relation is in some cases satisfactory, but in others, the analysis of the induced hydrogeological processes becomes essential to understand the functioning of the instability phenomenon and to predict the movements. Our project aims at introducing this missing link to allow a better understanding of the hydrogeological processes caused by particular hydroclimatologic conditions. It is a matter of determining the relation between hydroclimatologic phenomena and a critical hydrogeological state, which can lead a phenomenon of instability to reactivation or to a state of rupture. Unfortunately, this relation cannot be approached in the same way for all instabilities throughout the world. Owing to the specificity of each slope, the relation must be established individually, which may lead to a hydrogeological typology of unstable slopes. Accordingly, we propose a methodology comprising a succession of stages that are useful for the understanding of the hydrogeological processes underlying slopes. This approach aims at establishing a hydrogeological assessment on the scale of the slope or, if necessary, on a regional scale, in order to determine groundwater infiltration zones, trajectory and velocity of groundwater run-off as well as outlet areas. The usual ways of doing this are generally the measurement of the river flow rate, a pluviometric follow-up or direct hydrogeological measurements via piezometers. These traditional approaches generally remain limited by problems of space representativeness or life span; on top of this come problems of cost due to the often chaotic morphology of the slope instabilities and the fact that they are not easily accessible. To make up for these various problems, we propose a complete methodological approach to improve the understanding of the groundwater run-off systems for the unstable slopes: A characterization of the geological domain based on geophysical measurements in addition to the borehole data. This approach, original in the specific context of unstable slopes, makes it possible to obtain a three-dimensional representation of the heterogeneity which composes the subsurface grounds in a fast, efficient and economic way; A spatialization of the climatologic observations in order to take into account the distribution of the precipitations while considering the various states (rain and/or snow) that constitute the water contribution to a slope in relation to altitude; A hydrogeochemical characterization of the outlets, which makes it possible to assess the general groundwater circulation system within the slope, and to come to a decision about one or several possible trajectories of groundwater run-off; An isotopic characterization of the outlets, which allows estimating the altitude of infiltration of the groundwater. This characterization can also give some information on the infiltration zones. These methods also make it possible to characterize the general groundwater run-off system. In some cases, it is also possible to come to a conclusion, quantitatively speaking, concerning the mixing between several reservoirs and to estimate the time of residence of the water in the aquiferous system; A qualitative validation of the assumptions on the basis of artificial tracing tests which will attest the existence of hydrogeological connections between a catchment area and an outlet. This approach will also inform about the time of residence of water; A validation of the conceptual model on the basis of a bi- or three-dimensional numerical simulation of the groundwater run-off in order to validate the hydrological and hydrogeological relations. The calibration of these numerical models is based on a set of hydrogeological data, in particular the outlet discharge, the isotopic composition of water and/or their hydrochemical composition. These strictly hydrogeological numerical simulations can then be translated into a mechanical state to evaluate the stability of the slope. The whole of the points suggested by this methodological approach is not always essential to determine the hydrogeological processes that govern each slope. According to the characteristics of each study site and the means at disposal, some could be treated only partly, or not at all, depending on the established objectives. In the end, a hydrogeological typology of the unstable slopes should make it possible to propose a standard methodological approach and specific tools for analysis in direct relationship to the hydrogeological context which characterizes each slope. The methodological approach suggested in this PhD thesis was tested and validated on three representative sites through Switzerland and the alpine arc that present quite distinct characteristics of hydrogeological functioning. They are the slopes of Hohberg in the canton of Fribourg, Triesenberg in the Principality of Liechtenstein and Peney in the canton of Geneva. This validation phase allowed to assess the hydrogeological functioning of the three studied slopes and to evaluate the relation between the hydrogeological processes and the phenomena of instability. An approach such as that presented in this study opens up prospects to carry out means of sustainable stabilization with an aim of solving problems of slope instabilities.