We present results of sequential data assimilation of altimetric data in a linear tropical Atlantic model, using an analysis technique that we have developed. We use sea level anomaly fields obtained from altimetric data of the satellite Geosat. Despite the weak amplitudes of the anomaly fields, we obtain marked improvements of the model simulations due to assimilations. Thus, we show that altimetric data may be very useful for improving model simulations when adapted assimilation techniques are applied.