Soil management in crops is mainly based on intensive tillage operations, which have a great relevancy in terms of increase of atmospheric CO2, desertification, erosion and land degradation. Due to these negative environmental impacts, the European Union only subsidizes cropping systems which require the implementation of certain no-tillage systems and agro-environmental measures, such as keeping the winter cereal residues and non-burning of stubble to reduce erosion, and to increase the organic matter, the fertility of soils and the crop production. Nowadays, the follow-up of these agrarian policy actions is achieved by ground visits to sample targeted farms; however, this procedure is time-consuming and very expensive. To improve this control procedure, a study of the accuracy performance of several classification methods has been examined to verify if remote sensing can offer the ability to efficiently identify crops and their agro-environmental measures in a typical agricultural Mediterranean area of dry conditions. Five supervised classification methods based on different decision rule routines, Parallelepiped (P), Minimum Distance (MD), Mahalanobis Classifier Distance (MC), Spectral Angle Mapper (SAM), and Maximum Likelihood (ML), were examined to determine the most suitable classification algorithm for the identification of agro-environmental measures such as winter cereal stubble and burnt stubble areas and other land uses such as river side trees, vineyard, olive orchards, spring sown crops, roads and bare soil. An object segmentation of the satellite information was also added to compare the accuracy of the classification results of pixel and object as Minimum Information Unit (MIU). A multispectral QuickBird image taken in early summer was used to test these MIU and classification methods. The resulting classified images indicated that object-based analyses clearly outperformed pixel ones, yielding overall accuracies higher than 85% in most of the classifications. The choice of a classification method can markedly influence the accuracy of classification maps.