Remote sensing of significant wave height and sea surface wind speed from satellite altimeters is now conducted on a routine basis. Several years of experience with GEOSAT, ERS-1, ERS-2, TOPEX and POSEIDON, provided large data sets which are of particular interest for wave climate studies. For this purpose, long term validation is required because any few percent drift of the measurement can induce significant erroneous trend. In the present paper, differences between ERS-1, ERS-2, TOPEX and POSEIDON altimeter measurements are investigated over a 3-year time period. This is based on comparison of altimeter measurements at ground-track crossings. The analysis shows that long-term biases exist among these satellite altimeter measurements. The observed biases for backscatter coefficient and surface wind speed are low and can be taken into account to retrieve homogeneous wind data. For significant wave height measurements, almost time independent biases are observed between ERS-1, ERS-2 and POSEIDON. However, a singular behaviour is shown for TOPEX: the wave height measurements increase with time, about 0.45 m over the 3-year period, and as much as 0.25 m over the last year of data.