This article aims at introducing a spatial aspect of analyzing the relationship between economic growth and poverty, in which the geographical disparity plays an important role. A case study of growth poverty elasticity of Sumatra, Indonesia, using district-level data, is provided as an example of how spatial issues can be incorporated in the analysis. An application of spatial regression is introduced as a suitable method. The case study of Sumatra indicates that spatial influence exists in the poverty determinant model and this spatial relationship shall not be ignored in estimating GEP. Poverty incidence in Sumatra tends to be concentrated in particular locations which is influenced by spatial factors such as types of land, soil condition, transportation, and infrastructure. This suggests a poverty alleviation program should be developed targeting larger regional bases, rather than concentrating on various small areas. This method can be applied in a number of topics in agricultural economics as well as other fields relating to location factors and neighborhood influence.