A study about spatial variability of soils and their features is shown. This study is the initial stage of a work designed to appraise the use of organic residues to restore eroded areas and their use in order to improve soil structure and vegetation cover. The irregularity of mountain areas implies a very high variability of their environmental features. It provides necessary study of the spatial variation of soil parameters to delimit possible experimental areas and to find their background values. Because soil studies are destructive, we chose to do a soil sampling, on irregular mesh, from the locations we divided into plots to interpolate the soil parameters inside them. The study area is located in an eroded area sited on a mountain environment. It is on a simple slope between the summit and the footslope. Lithology in this area is shaped for an intercalation of marls and hard limestones of Eocene period. Temperature average is 11ºC and precipitation is 850mm/year but very irregular. The soil moisture regime is xeric; it is favored for the limited thickness of the soils.In spite of the limited studied surface, the soil typology is very varied. Variations in the parent material and landscape position are the main causes of this diversity. Usually the soils are thickness (less than 40cm). If the parent material is marl they are eroded soils, A-Cr type, eroded (Lithic Haploxeroll or Lithic xerorthents); and if the parent material is a hard limestone they are reddish very developed soils, A-B-R type (Lithic Haploxerept or Lithic Haploxeralf). In accumulation areas appear soils more thick and slightly evolved, A-B-C type (Typic Haploxerepts). Vegetation is very influenced by the human activity. At times in agriculture, at times in ovine cattle or in the forestry use of the little woods which have resisted the human pressure in the less productive areas, over hard limestones. The sampling area was set in a rectangular surface of 200x300m. It was sampled at two levels: four profiles of representative soils and 30 aleatory points. Each one of these points has been sampled to 0 to 8cm and 8 to 20cm, and the points over deep soils a third sample has been sampled to 20 to 60cm. The parameters and soil analyses studied were: location (UTM) and soil thickness, bulk density, aggregate stability, water retention 33kPa and 1500Kpa, Munsell soil color, gravels, soil particle size, pH, EC, % organic C, %N , %CaCO3 equivalent. Other features evaluated indirectly were: C/N, erodibility factor (K), AWC (Available Water Capacity) and total organic C and N (kg/m2). The GIS program Idrisi Kilimanjaro v14.0 was used to calculate the digital elevation model (DEM) and to do kriging and other surface analyses. 3D-Maps of soil cover show a good correlation with the environmental parameters and allow establishing models of distribution patterns of the diverse characteristics analyzed with geology and the topography. The zones with less organic C are located in the eroded zones and with a greater slope; the CaCO3 is minimum in the soils more evolved and located in the high zone of the landscape and maximum in the eroded zones contain more superficial parental material. A very detailed soil map was made, and correlates well with observed reality. Another interesting product is to quantify the absolute values of some important soil parameters for the environment like the total content of organic C and the capacity of water storage of the zone. This derived the study of erodibility or special correlation between some soil parameters and the structural stability can be a useful tool to understand dynamics of the system. This shows some questions about soil spatial variability and its implication in some of the models. The random sampling avoids overvaluing some soils that are not abundant. A deep knowledge of the genesis and the typology of soils - directed sampling - is necessary for a correct interpretation of the results of the random samplings,especially when these introduce an important statistical manipulation. The combination of the GIS, and its tools, with the soil random samplings, together with the study of its apparent density, allows the making of a simple detailed soil cartography and quantification of the parameters of it: This increases its practical interest remarkably and it approximates it to the reality. The accomplishment of these models can be the base of the study of the dynamics of soils, its different phases and from its interaction with the environment. At the same time, it implies a revision of the valuation that is made by means of models based on directed samplings and to small scales.