The multi-scale representation model of complex landform based on two-dimensional wavelet transformation is proposed in this paper. The new idea and principle of this model can completely get rid of intersection of different contours which commonly occurred in almost all models used before. Thus

it could be said that by theory and method the model can preserve the veins of landform

maintain the coordinate relations among the contour lines

besides

it has adaptive property. The experiments on that model give the following conclusions:① The landform characteristics such as the peaks

saddle points

valley lines

intersections of valleys

ridge lines of mountains can be preserved very well

so the structure and roughness of the global landform will be retained.② The relative positions of the contour lines are maintained precisely

so the shape of the landform such as the outines of mountains

the directions of ridges and valleys are retained quite well.③ The rates of changes of the amount of information

the simplification of the contour lines

the extent of landform generalization for different levels of decomposition or reconstruction are consistent and stable

besides

they coincide with the visual feeling and standard of the ordinary people.④ The Lipschitz index

which is related to the roughness of the landform

is used to calculate the threshold value in the model

thus the threshold which control the amount of information being retained is directly linked with the complexity of the landform and could be adjusted adaptively.⑤ Because of the flexibility of the threshold value and the stability of the method

the multi-scale representation based on that two dimensional wavelet model is suitable for more kinds of landforms.