Stem cell therapies have shown promise in cerebral palsy (CP). However, their effects on brain metabolites, measured by proton magnetic resonance spectroscopy (¹HMRS), remain unexplored. In this randomized clinical trial, we evaluated ¹HMRS measures (N-acetyl aspartate [NAA], choline [Cho], creatine [Cr], myo-inositol [mI], NAA/Cho, and NAA/Cr) within periventricular white matter (PVWM) in children with CP at baseline and 12 months after a single intrathecal injection of umbilical cord-mesenchymal stem cells (UC-MSCs:20 × 10⁶) or umbilical cord blood-mononuclear cells (UCB-MNCs:5 × 10⁶/kg). Generalized Estimating Equations (GEEs) were employed to assess treatment efficacy, the adjusted effects of sex, CP type, and gestational age (GA) on post-treatment findings, and the association between metabolites and gross motor function measure (GMFM-66). Seventy-three participants were included: UCB-MNC (n = 27), UC-MSC (n = 26), and sham (n = 20). Primary analyses indicated no significant time*treatment interaction effects for any of the metabolites. In exploratory analyses, a significant CP type*treatment interaction was found for the post-intervention NAA/Cho (UC-MSC vs. sham; P-value = 0.02). Significant GA status*treatment interactions were also observed for post-intervention Cho (UC-MSC vs. sham; P-value = 0.009) and post-intervention mI (UCB-MNC vs. sham; P-value = 0.03). Additionally, longitudinal increases in NAA/Cr and NAA/Cho were positively associated with GMFM-66 improvement during the follow-up in UCB-MNC and UC-MSC groups, respectively, whereas Cho*time interaction was linked to smaller functional gains in the UCB-MNC group over time. Neither UC-MSC nor UCB-MNC had a significant overall effect on measured metabolite concentrations/ratios within the PVWM at 12 months following a single injection. Exploratory findings should be interpreted cautiously.