Digital Elevation Model (DEM) interpolation is one of basic functions for spatial description and spatial analysis in GIS
and related spatial information fields. Interpolation can be viewed as a function for estimating the heights of unknown
points using a set of proper known data. It is a key problem of DEM. Inverse distance weighting (IDW) interpolation is
the most commonly used in DEM. The reference points selected by IDW might not be well distributed in space. This
leads to the discontinuity problem of interpolated DEM surface and some artifacts might be generated. In order to solve
the problem caused by ill-distribution of the number and position of reference points in searching process, this paper put
forward a new surface interpolation model about first-order natural neighbor interpolation. The use of first-order natural
neighbor interpolation based on TIN can adapt well to poor data distributions because inserting into a point generates a
well-defined set of neighbors. In the fitting process, according to range of influence composed by first-order natural
neighbor points and the triangle area as weight base of the known point, a non-linear fitting equation can be constructed.
Comparative experiments show that this method has higher precision and more practical application value.