Cross‐syndrome comparisons offer an important window onto understanding heterogeneity in mathematical learning disabilities or dyscalculia. The present study therefore investigated symbolic numerical magnitude processing in two genetic syndromes that are both characterized by mathematical learning disabilities: Turner syndrome and 22q11.2 deletion syndrome (22q11DS). We further verified whether the phenotypic outcomes of these syndromes emerged from the same or different cognitive processes and therefore examined whether numerical impairments were related to working memory deficits, often observed in these syndromes. Participants were 24 girls with Turner syndrome, 25 children with 22q11DS and 48 well‐matched typically developing control children. All children completed a symbolic numerical magnitude comparison task and four additional working memory tasks. Both groups of children with genetic syndromes showed similar impairments in symbolic numerical magnitude processing compared to typically developing controls. Importantly, in Turner syndrome, group differences in symbolic numerical magnitude processing disappeared when their difficulties in visual‐spatial working memory were taken into account. In contrast, the difficulties in 22q11DS were not explained by poor visual‐spatial working memory. These data suggest that different factors underlie the symbolic numerical magnitude processing impairments in both patient groups with mathematical learning disabilities and highlight the value of cross‐syndrome comparisons for understanding different pathways to mathematical learning disabilities or dyscalculia. To understand heterogeneity in dyscalculia, we investigated symbolic magnitude processing in two distinct genetic syndromes that are both characterized by impairments in mathematics, Turner Syndrome and 22q11 Deletion Syndrome. Both patient groups showed similar deficits in symbolic magnitude processing, which were explained by poor visual‐spatial working memory in Turner Syndrome but not in 22q11 Deletion Syndrome. These data show that different cognitive deficits underlie phenotypically similar impairments in symbolic magnitude processing and highlight the need to consider different pathways to dyscalculia.