According as general houses' prices data, this paper, based on spatial analysis function of Geographic information system(GIS), using semi-variogram of spatial statistics, studies spatial heterogeneity of general houses' prices distribution in Dongguan quantitatively. The results from the analysis indicate: general houses' prices have both spatial autocorrelation and sometime local spatial heterogeneity, it can be found that the spatial distribution of general houses' prices takes on a zonal anisotropy by anisotropic variability analysis, which means that there are different structural
characteristics in different directions for general houses' prices distribution; isotropic variability analysis reveals that: the
semi-variogram of general houses' prices distribution in Dongguan is best described by spherical model, changes of general houses' prices distribution are affected by both structural and random factors; the ratio of random variance (nugget) to total variance(sill) is 37.5%, therefore the spatial correlation of general houses' prices is a kind of medium correlation with Nugget/Sill being between 25% - 75%, its spatial correlation range is 16.62 kilometres; the ratio of
structure variance(partial sill) to total variance is higher than the ratio of random variance to total variance, this means that certain factors' contributions to the spatial variability of houses' prices is more than random factors' contributions.