The visuo-spatial sketchpad (VSSP) is a temporary storage system for generating and manipulating visuo-spatial images. Preserving and manipulating limited on-line visual information is an important function of the visual processing system. Some recent findings have confirmed that the VSSP may be divided into two sub-components processing static or dynamic visual information. This study adopted spatial localization tasks, as more typical spatial working memory tasks, to explore the representation construction characteristics of the static and dynamic VSSP as well as its individual differences. Pilot group and non-pilot group were recruited to make a compare of their processing properties in the static and dynamic spatial localization tasks. Experiment 1a set up two single-target localization tasks in static spatial version, namely distance reference spatial localization tasks and no-distance reference spatial localization tasks. Experiment 1b was a variant of Experiment 1a with double-target localization tasks under each conditions. Experiment 2a developed two dynamic spatial single-target localization tasks, as trajectory reference spatial localization tasks and no-trajectory reference spatial localization tasks，respectively. Experiment 2b extended two double-target localization tasks on the basis of Experiment 2a. The experimental design focused on effects of the reference condition, the target amount and the spatial nature on the spatial localization. The results showed that in static display the distance reference spatial localization tasks were easier tasks and the no-distance reference localization tasks were more difficult. There was no significant difference in the distance reference localization tasks between pilot group and non-pilot group. However, pilot group performed better in the no-distance reference localization tasks. In dynamic tasks, the trajectory reference localization tasks were easier tasks and the no-trajectory reference localization tasks, which required immediate construction of trajectory representation, belonged to more difficult tasks. Pilot group had higher accuracy in both dynamic tasks. Compared to the condition of trajectory reference, the condition of no-trajectory reference caused non-pilot group score significantly lower. The difference between two groups in the double-target localization tasks was much more greater than which in the single-target tasks. Compared to the control group, pilot group exhibited double target synergies in the double-target tasks. This study has found that pilots’ VSSP had better performance on processing visual-spatial information, involving supervisory attentional system engaged in computation and reconstruction spatial representation. Pilot could complete clear distance representation more precisely in static spatial display while their computational advantage in dynamic tasks was reflected in constructing on-line trajectory representation and then forming a stationary configuration. These findings raise questions about the hypothesis that the VSSP can manipulate complex visual-spatial information. The innovations of this study are as follows. First, dynamic spatial localization tasks were developed according to Gordon’s static tasks(1986). Second, double-target localization tasks were set up under the static and the dynamic spatial conditions, and the synergistic effect of the double targets was found better in pilot group. Third, two kinds of control tasks were adopted to comparatively explore the representation construction of VSSP in static and dynamic localization tasks. Finally, experience advantage of pilot found in this study provides empirical reference for the future selection and training of pilot.