Geospatial Information systems (GISs) play an active role in decision-making processes in many disciplines, research, and/or management using spatial data, at the global, regional, local or municipal level. More effective use of GIS, however, requires explicit knowledge of the uncertainty inherent in the data, particularly positional data. This becomes more important as the potential impact of decisions based upon GIS increases. The widespread use of GIS as a decision support system is, therefore, dependent on the development of formal modeling of uncertainty in spatial databases. This paper focuses on positional errors that refer to the discrepancy between the true location of certain geometric primitives and their measured or expressed locations. If true values are not available, uncertainty is substituted for error. The current uncertainty models of geometric elements are critically evaluated using analytical methods. Particular emphasis is placed on modeling uncertainty of line objects. The result of a case study on uncertainty assessment showed that magnitude of modeling uncertainty versus measurement uncertainty can be considerable which is usually ignored in GIS analysis.