We tested the two-process model of spatial localization (Adam, Ketelaars, Kingma, & Hoek, 1993) by measuring accuracy and precision of localization performance in three stimulus conditions: (i) no-saccade — observers were to maintain gaze on the fixation mark while target appeared for 1 s, (ii) saccade to on-target — they had to quickly saccade to and fixate on target remaining on for 1 s, and (iii) saccade to off-target — they had to quickly saccade to the target appeared for 150 ms. Observers’ task was to adjust position of a mouse cursor manually so that its position corresponds to the perceived position of the target. Results with six observers are consistent with the model suggesting localization performance to be mediated by memory-guided saccade for brief target and visually guided saccade for long duration target. The results further suggest that the memory-guided saccade provides relatively less accurate and precise information for localization and is a function of target eccentricity.