Abstract:The soils in intertidal zone in Jiangsu Province are formed and distributed under their unique conditions. They vary with vegetation and water conditions, which results in unequal multispectral characters Therefore, the types of soil can be recognized by computer image processing using LANDSAT CCT. This paper deals with the comparison for different methods for image processing including ratio processing, supervised and unsupervised classification, etc. used in recognition of soil types in intertidal zone of coastal region giangsu.The experiment showed that selected ratio was hater than sequenced ratio, supervised classification was better than unsupervised one. Among supervised classifications, the method of 13 training-land-classes was suitable, by which 3 sub-groups, i.e meadow coastal saline bog coastal saline soils and wet beach coastal saline soils; 4 genera according to different teztures and 7 species according to the degree of salinzation were recognized and classified under the soil great groups of coastal saline soils. Furthermore, the accuracy of the computer classification was checked with field actual measured data obtained from long-term experimental sites installed in the experiment region, which showed a mean relative error of 17.6%.