Survey spatial representations are map-like mental representations in which directional relationships among landmarks are preserved. Survey representations allow people not only to re-trace routes already experienced, but also to find new routes and shortcuts (Golledge et al., 1999). This study investigated whether and to what extent verbal and spatial working memory (WM) are implicated in the construction of survey representations. We adopted a dual-task paradigm, asking participants to learn a new environment from navigation and, concurrently, to perform either a verbal or a spatial task, assumed to load verbal and spatial WM, respectively. Ninety undergraduates were assigned to one of three groups according to concurrent task condition: articulatory suppression, spatial tapping, or control (no concurrent task). Acquisition of a survey representation was tested by asking participants to perform direction estimations and shortcut tasks. The results supported the involvement of spatial WM in the acquisition of survey knowledge, showing significant differences between the spatial tapping group and control group for the survey measures.