In order to better reproduction of the information that is as close as possible to reality, we have established a procedure for analyzing images based on visual interpretation and digital processing. In digital processing, supervised classification by maximum likelihood, based on a fieldwork allowed to discriminate elements of land. Recovery of geometric shapes by scanning fallow and bare soil from panchromatic Spot image yielded the map of agrarian structures. Different queries and statistical operations performed were used to assess the areas of agrarian structures. The analysis of the dynamics of agrarian structures showed that the period 1986 - 2000 was marked by environmental degradation, while the period 2000 - 2004 has seen a regeneration of the natural environment. Treatments applied to images and ensuing results indicate the feasibility of monitoring by remote sensing of agricultural structures in northeastern area of Côte d'Ivoire. This research shows that the Geographic Information System is an effective tool for the evaluation of agrarian structures. However, manual scanning of elements in the environment requires careful attention to avoid skewing the results.