This paper proposed a framework potential for assessing regional spatial and temporal variations of air pollution exposure risk by combing Geo-information Technology Aided Proximity Model (GTAPM) and GIS interpolation technology in areas with limited access to data inputs (e.g. emission data, meteorological data). In this process, the exposure risks to sulfur dioxide (SO₂) for Dallas and Ellis counties in 1996 and 2002 were assessed firstly at annual scale based on GTAPM; then, the spatial and temporal variations of annual SO₂ exposure risk in Dallas and Ellis counties were evaluated based on GIS spatial interpolation technology and overlay analysis operation. The results indicate that the annual SO₂ exposure risk in Dallas and Ellis counties are spatially different in both 1996 and 2002. While areas with lower levels of variations were mainly distributed in most central parts of study area, the areas with relative higher levels of variations emerged in the northern part of Dallas county and southeastern of Ellis County. The results also suggest that while the annual SO₂ exposure risk in minority areas with relative high exposure risks in 1996 seemed to be reduced over the period of 1996 to 2002, those in majority areas with relative lower exposure risks in 1996 increased a lot. The annual SO₂ exposure risk of entire study area appeared to increase. Therefore, it can be concluded that GTAPM can effectively and efficiently evaluate air pollution exposure risk and its spatial and temporal variations in area without air quality observation data and related input data for air dispersion modeling.