The cartography and quantification of irrigated fields in the context of decreasing rainfall constitute a key element for water resources management. Therefore, in this context, the use of remote sensing methods applied to Landsat-type images with a high spatial resolution for monitoring the changes in land use in general and irrigated crops, in particular, is highly relevant. This paper aims to present a method for mapping spatial and temporal changes in irrigated parcels in the Guigou Plain, located in the central Middle Atlas, based on Landsat images and fieldwork. For the years 1985, 1998, 2010 and 2018, the use of a supervised classification method based on the principle of machine learning, fed by precise field surveys, has made it possible to highlight a significant extension of irrigated areas to the expense of pastureland and rainfed crops. Over the entire period under consideration, the results obtained with good precision (98.5% overall accuracy) showed that the area under irrigated crops has increased from approximately 699 ha to 3988 ha, i.e. an increase of 570%. The corollary of this increase is strong pressure on the water resource, especially groundwater. This information on the total extension of irrigated plots can be taken as a reference in the perspective of reasoned management of water resources in the sector.