Previous investigations into working memory impairment in autism have yielded mixed findings. The present study is a continuation of research from a previous investigation (McGonigle, Tait & Smith, 2005) and examines visuospatial working memory in a sample of children with “low-functioning” autism (N=12), relative to both a clinical control group (N=8) diagnosed with Non-specific developmental delay (NSDD) and a typically developing control group (N=7). Both control groups were age-matched and the clinical control group was also matched on nonverbal IQ. Participants were tested using a newly developed, computer-based task, the ‘v-Music 11-12’. The typically developing control group performed significantly better than the clinical control groups on all three measures (Highest level attained, Average trials per level and Regression error proportion). There were no significant group differences between the target group and clinical control group, but the autism group performance was impaired relative to the NSDD group on all three measures and it should be noted that sample sizes were small. HLA did not correlate significantly with age in either clinical group and correlated with nonverbal IQ only in the NSDD group. Evidence of learning subgroups was not replicated (McGonigle et al., 2005), but the majority of participants with NSDD reached level 8 or above (most successful subgroup), whereas this was not the case in the autism group. It cannot be concluded that visuospatial working memory impairment in “low-functioning” autism is different in either degree or type from the impairment found in NSDD. Further investigations into working memory in “low-functioning” autism, employing appropriate comparison groups, are required.