Abstract: By comparing with current methods,we propose forward a new urban land classification model based on k-order data field point to point-factors.This model has three key techniques links: on the layer's classification,the model provides more flexible policy and expresses the spatial distribution correlation of different class factors.Moreover,we desert the old classification method purely according to property.On the effect area division,we build k-factor influencing environmental effect area which has an advantage in no crack and no cover by k-order Voronoi polygon nearest neighborhood division.With k-factor influencing effect area,we can control model precision and meet the needs of Urban Land Classification in complicated conditions.On the calculation of assessing point effect value,we present a factor value calculating strategy corresponding to k-factor influencing environment,especially in three kinds of approximation as well as more kinds of multi-factor calculating rules.