This study aims to develop several spatial sampling methods to check with TEDS data. This spatial technique is considered valuable for further research for several reasons. First, it is a way of optimal design for existing samples. Second, it avoids spatial autocorrelation and spatial heterogeneity to produce a better measurement. This study applies voting data from the 2012 presidential election and 2010 mayoral election of TEDS to test validity. As expected, the two-dimension spatial sampling method is a more accurate estimation for election studies.