In the present study, the effect of phonological and working memory mechanisms involved in spelling Italian single words was explored in two groups of children matched for grade level: a group of normally hearing children and a group of pre-verbally deaf children, with severe-to-profound hearing loss. Three-syllable and four-syllable familiar words were presented to the two groups for spelling to dictation. Three conditions were used: simple spelling, concurrent articulation, and foot tapping. Verbal digit span was also assessed. Overall, the performance of deaf children tended to be lower compared to hearing children, but not significantly so. Concurrent articulation produced more errors than tapping in both groups. Regression analyses showed that the main predictor in all three tasks was school level, however the proportion of variance explained by this factor was much greater in the dual tasks, in particular in concurrent articulation. Qualitative analyses of errors showed a worse performance of deaf children, with a greater proportion of mixed errors compared to hearing children. They also showed a greater proportion of phonologically plausible errors compared to hearing children, presumably due to their deprived auditory representation, and/or to phonological representations that rely to a large extent on lip reading and kinesthetic and visual perception of articulatory gestures.