Abstract: SAR remote sensing, with its advantages of cloud penetration, all-weather coverage, high spatial resolution, day/night acquisitions, and signal independence of the solar illumination angle, prevails for far-ranging application in many fields. In present study, based on the characteristic of the backscatter signatures of various targets as a function of time, multi-temporal ERS-2 SAR images were applied to classify agricultural land use in the often-cloudy costal tideland reclamation region of Hangzhou Gulf, Zhejiang Province. After pre-processing the multi-temporal SAR images, different classification methods corresponding to the different spatial zones of the study areas were carried out. The study results show that SAR remote sensing, instead of multi-optical remote sensing, has great potential in the agricultural land use investigation.