The World Health Organisation (WHO) recommends rapid molecular diagnostics to improve bacteriological confirmation of tuberculosis in children. TrueNat MTB, MTB Plus and MTB-RIF Dx assays (Molbio Diagnostics, India), recommended by WHO, hold potential as point-of-care tests in resource-limited settings. Using stool samples with these assays could enhance testing access, improve linkage to care, reduce costs, and increase cost-effectiveness over traditional methods. However, evidence on their costs and cost-effectiveness is limited and needed for informed policy decisions on adoption and scale up. We used a decision-tree analytic modelling approach, time-and-motion study, and routine data to estimate the potential impact of implementing stool-based TrueNat testing for the diagnosis of pulmonary tuberculosis in children within Nigerian primary healthcare settings on healthcare outcomes, resource use, health system costs, and cost-effectiveness relative to the standard of care (SoC). The cost per test was $13.06 (standard deviation; $0.77) for TrueNat and $16.25 (standard deviation; $1.34) for Xpert. For every 100 children with presumptive tuberculosis, the stool-based TrueNat testing intervention was projected to increase case detection rate by 29% (95% uncertainty interval [UI 9-48]) and bacteriological confirmation by 6% (95% UI 3-8). Diagnoses at primary health centres (PHC) would increase by 34% (95% UI 20-46), averting 7 (95% UI 2-12) deaths and 182 (95% UI 59-311) discounted DALYs. Although resource use and health system costs increased by $6,172 (95% UI -2,427-10,455) per 100 children, the incremental cost-effectiveness ratio of $34 per DALY averted suggests cost-effectiveness at thresholds of 0.5×GDP per capita. Implementing stool-based TrueNat testing has potential to increase access and reduce direct health system costs associated with the diagnosis of pulmonary tuberculosis in children in routine health care settings. Such an approach is likely to represent a good value for money compared to SoC.