LANDSCAPE ANALYSIS AND CLASSIFICATION METHODOLOGIES. THE EXAMPLE OF THE ESTRELA PROJECT. Landscape ecology studies experience at the moment an increase of interest by researchers, developers and end-users. However, the subject matter is still too dispersed, not only in methods used but also in objects of analysis as well as space and time-scales. Progress on quantitative methods and the use of Geographical Information Systems frequently lead to excessive emphasis on physiognomic patterns to the detriment of landscape dynamics. Furthermore, the study of landscape as a geosystem suffers from problems related to calculation of mass and energy fluxes, which are barely possible, even with the powerful and costly equipment available only in few laboratories. The integrated environmental research, albeit at a more modest level, requires a multidisciplinary approach (defended by many, but seldom applied). The objective of the ESTRELA Project is to study the relations between the physiognomy and dynamics of the landscape in a mountain environment. Geomorphology, climate and vegetation are analysed at three spatial scales. The research carried out on each domain is presented and an example is given of their integration in a sample-site. The conclusion refers to the temporal transformations of the landscape of the Serra da Estrela. The problem of landscape units definition is still under investigation.