Good water quality is essential for all humans, animals and plants. In Europe the Water Framework Directive (WFD) aims to protect and restore water bodies in order to a reach good status, and to prevent deterioration. The Water Monitor Department of the Flanders Environment Agency (VMM) has e.g the task of measuring, modelling and reporting surface water quality in the framework of the WFD. Therefore, Chlorophyll-a (Chl-a) concentration is currently measured in 100 rivers and 16 lakes in Flanders. Chl-a is typically considered as a proxy for the amount of biomass present in the water. An excessive amount of Chl-a might point to the presence of harmful algae blooms. Thanks to the Copernicus programme and Sentinel-2A and B, an archive of free and open data is available since July 2015. This archive can be used to derive water quality variables like Chl-a spatially and temporally. To support VMM, a processing workflow was developed to derive Chl-a in near-real time from Sentinel-2 data. An atmospheric and adjacency correction with iCOR (De Keukelaere et al., 2018) was used for obtaining water-leaving reflectance. Existing Chl-a retrieval algorithms were assessed for application over Belgian surface waters. A selection of blue-green ratio, red-nir ratio and combined algorithms were considered as well as a Mixture Density Network (MDN) (Pahlevan et al. 2020), a machine learning approach, trained on a global in situ dataset. Based on the assessment, MDN was selected for implementation in the workflow. Chl-a products are accessible in Terrascope, the Belgian Copernicus collaborative ground segment, but also via a user-defined WaterMonitor viewer including in situ Chl-a data for easy comparison with Sentinel-2 based Chl-a. A similar NRT service demonstration but extended with other water quality variables like turbidity and suspended particulate matter was setup for Vietnam. A co-creation trajectory for a similar service for the transboundary lakes Tanganyika and Kivu in East Africa is ongoing.