Spatial variability is the focus of geostatistics in GIS analysis. It is an important tool to analyze the spatial data values and their locations. Spatial variability present in the sample data is assessed in terms of distance and direction, and can be described as surfaces. Methods of GIS-based spatial variability calculation and risk assessment are introduced. As an example application, spatial variability’s of four factors for oil production in petroleum engineering are analyzed. With kriging techniques for interpolating surfaces, risk images of four factors are obtained. Order weights based on rank-order approximation and criterion weights based on analytical hierarchy program are calculated. Therefore, the comprehensive risk assessment results are worked out. Finally, the calculating results are analyzed, and some advice is proposed.