Marine litter detection based on satellite remote sensing is becoming a reality but the current state of the art is not mature. The main activities being developed in that field are based on the use of different multispectral bands (e,g, Sentinel-2) to detect a sign of object presence on the surface and then discard that these signals are not originated by vegetation. This is a very promising approach but, since the current activities are premature, and there is a lack of ground truth to compare with the remote sensing measurements, a method to establish the areas in which high concentration of marine litter, including microplastics and macroplastics, is needed. The proposed idea is to use the expertise available at Deep Blue Globe, a company offering artificial intelligence solutions based on space technology to the benefit of the maritime industry, to simulate and perform a plastic concentration forecast which would then be studied with more details thanks to the multispectral data from Sentinel-2. The proposal consists of the following steps: 1) Acquire and process Sentinel-3 raw telemetry (level 0) and Copernicus Marine Service forecast (level 4). 2) Use Deep Blue Globe algorithms about sea surface wind, currents, waves and swell to compute the paths of the marine litter in open ocean. 3) Identification of the “sink” where they would concentrate (note that, as buoys, litter will tend to concentrate in specific areas of the ocean). 4) Acquisition of Sentinel-2 multispectral bands to identify signal of objects on the sea surface 5) Verification with ground truth making use of existing agreement between Deep Blue Globe and vessel navigating worldwide that is recollecting water samples to determine plastic concentrations (note that existing agreement referred to the scope of the capabilities from Deep Blue Globe to perform maritime route optimization and it was not related at all about marine litter activities).