Different sampling strategies have been tested for estimation of phytoplanktonic biomasses in a bay submitted to estuarine conditions and high velocity of currents. Daily variations showed that for a precision of 5% of the daily mean, it is necessary to sample with a 1 hour and a half time scale. Systematic sampling appears to be more preceise than the random sampling. A sampling during 35 days at low tide and at high tide, during the phytoplanktonic bloom, showed the large variability between the days. A precision of 30% of the mean biomass of the month is obtained with a 8 days time scale. Comparison between tidal cycle sampling (10% of precision) with a punctual sample in the middle of the day did not show large differences. As the main variability is between the days and not during on day, the autors suggest to sample 2 daus by week during the bloom. Different sampling strategies have been tested for the estimation of phytoplanktonic biomasses in a microtidal bay submitted to estuarine conditions and high velocity of currents. The study has been realised in two areas in the center of the bay (station 1, fig.1) wich is well representative of the evolution of the whole bay (Héral et al., 1983) and in the external estuary of the river Charente (station 2, fig. 1)