Background: Children with chromosome 22q11.2 deletion syndrome (22q11.2DS) exhibit nonverbal learning disability that may manifest in part because of working memory (WM) deficits. 22q11.2DS is a complex developmental disorder with serious physical, learning, cognitive, and psychiatric symptoms including a risk of developing schizophrenia 30 times that of the general population. WM impairment likely contributes to and exacerbates learning difficulties, school problems, existing neuropsychological disorders such as attention deficit hyperactivity disorder (ADHD); and a poor WM may be a biological risk marker for future mental illness. WM impairment is established in this population, but less is known about its neurological origins. Frontoparietal cortical development and function are key to WM processing. In the neurotypical developing brain, studies indicate activation associated with WM shifts from parietal to frontal regions with age. However, in children with 22q11.2DS, activation is restricted to the frontal cortex, and volumes are reduced in parietal regions where abnormal tractography abides. The overarching aim of this study was to determine the neural origins of WM impairment in people with 22q11.2DS. Methods: We measured WM in children and adolescents with (n = 29) and without (n = 27) 22q11.2DS using the WISC-IV and a computer-based spatial working memory task (SWMT) task. Participants’ brains were scanned using high-resolution magnetic resonance imaging (MRI) and diffusion tensor imaging (DTI). Focusing on brain morphometry and structural connectivity within frontoparietal networks, we investigated neural underpinnings of WM processing in 22 children with 22q11.2DS and 19 typically developing (TD) controls ages 7 to 16 (M = 12.13 ± 2.41). A connectome mapping network involved in WM processing was constructed by superimposing cortical segmentations on white-matter tractography. Results: Children with 22q11.2DS had impaired working memory performance. Individuals’ performance on our SWMT moderated the association between diagnosis and gray and white matter macro and microstructure. Children with 22q11.2DS with better working memory had larger lateral orbitofrontal volumes, greater axial diffusivity in the left superior frontal to superior parietal tract, and smaller volume in the right superior frontal to lateral orbitofrontal tract. Poorer performance in children with 22q11.2DS was associated with smaller right superior parietal and superior frontal cortical volumes. Conclusions: Children with 22q11.2DS performed worse on measures of working memory. Their performance was related to regional cortical volume differences and white matter microstructure abnormalities in the frontal and parietal lobes. These are brain regions consistently implicated in WM processing.