Background: The role of gender in both spatial and mathematics performance has been extensively studied separately, with a male advantage often found in spatial tasks and mathematics from adolescence. Spatial reasoning is consistently linked to mathematics proficiency, yet despite this, little research has investigated the role of spatial orientation and gender in the relationship between spatial reasoning and mathematics. Aims: In the present study, three spatial reasoning constructs (mental rotation, spatial visualization, and spatial orientation) were examined for their unique contributions to mathematics performance in two samples (Study 1: grade 5; Study 2: grade 8). In light of the emerging gender gap in mathematics as children develop, these relationships were explored as a function of gender. Sample: Eighty-four fifth-grade students participated in Study 1 (43 females, 41 males; mean age = 11.19 years). Nine hundred and three eighth-grade students participated in Study 2 (498 females, 405 males; mean age = 13.83 years). Methods: The three spatial reasoning constructs (mental rotation, spatial visualization, and spatial orientation) were examined for their unique contributions to mathematics performance for females and males in general and across different mathematical content (geometry–measurement and number sense). Results: Spatial factors accounted for 51% of the variance in math scores in Study 1 (grade 5) and 32% of the variance in math scores in Study 2 (grade 8). In both studies, spatial factors predicted a larger proportion of variance in geometry–measurement than for number sense. Spatial orientation was found to be a unique contributor in all mathematics models, object-based spatial skills (mental rotation and spatial visualization) varied in their contribution to math performance depending on mathematics content and gender. Conclusions: The present work highlights the unique contribution of spatial orientation in the spatial–mathematics relationship and provides insights into the nature of gender differences in mathematical problem-solving as a function of spatial reasoning and mathematics content.