The current satellite imagery management platforms are useful tools, but they assign labels with fix priorities to each patch of a satellite image based on the contents of the image. These tags with fixed priorities do not respond to all needs since each user is different and the uses given to the satellite images depend on the work carried out. Faced with this, we propose a new paradigm in which the priorities of the labels change according to the user who makes the request for the satellite image. For this purpose, we use machine learning and artificial intelligence techniques. With this proposal, one could have a platform that will give a better result at the time of search and manage satellite imagery.