[en] Spring Phaeocystis globosa blooms in the Belgian Coastal Zone (BCZ) are now recognized to be related with both North Atlantic Oscillation and anthropogenic eutrophication. In this dynamic coastal zone, mechanisms causing such blooms are not totally understood yet, especially variabilities in diatoms - Phaeocystis successions. The choice of monitoring stations is thus very important for studying eutrophication impacts in areas like BCZ where geographical extend of river loads is highly variable but, manual processing of discrete samples is time consuming and limits the number of stations sampled. This traditional method is not suitable for rapid analysis of the high temporal and spatial distribution variability of plankton communities as required to calibrate numerical models. We propose a method of near-real time monitoring of plankton assemblages which combines automatic recognition by machine learning algorithms using the Zoo/PhytoImage software (http://www.sciviews.org/zooimage/index.html) and the FlowCAM (Flow Cytometer And Microscope). This method is currently developed for the AMORE III project (Advanced Modelling and Research on Eutrophication, http://www.ulb.ac.be/assoc/esa/AMORE/objectives.htm) and ultimately aims to detect and understand plankton successions and their spatial and temporal variations along continuous transects. The method provides abundances and size spectra of the entire sample or for each group in near-real time aboard the RV Belgica oceanographic ship.