Objectives Working memory (WM) capacity is similar for signed (SL) and spoken (SpL) language yet underlying temporal and spatial processing mechanisms may not be identical. To investigate this, two studies with deaf native signers (DS) and hearing non-signers (HN) were conducted. Methods DS and matched HN groups performed WM tasks with varying temporal and spatial demands in study 1 at encoding (temporal, spatial and mixed presentation styles) and in study 2 at retrieval (forward and backward span) and with abstract spatial demands (math span). Results DS performance was inferior with high temporal demands at encoding (temporal style) and retreival (forward span). There was no difference between groups with high spatial order demands at encoding (spatial style) or retrieval (backward span). DS performance was worse when abstract spatial processing was involved (math span). Conclusion WM processing mechanisms for SL and SpL differ for temporal information at encoding and retrieval and for abstract spatial information.