Objective Microbiological confirmation of tuberculosis (TB) in pediatric cases is challenging due to its paucibacillary nature and difficulty in specimen collection. This study aimed to validate stool as an alternative sample for the diagnosis of pediatric pulmonary TB via Xpert MTB/RIF (Xpert) assay. Materials and Methods This cross-sectional study included 75 pediatric patients up to 10 years of age with signs and symptoms suggestive of TB. From each recruited patient, pulmonary and stool samples were collected in a sterile container. The collected samples were subjected to Ziehl-Neelsen staining, BACTEC MGIT 960 culture (MGIT), Xpert, and in-house multiplex polymerase chain reaction for TB diagnosis. Results About 13.33% (10/75) of the pulmonary samples and, of them, 50% (5/75) of the stool samples were positive by Xpert assay. The sensitivity and specificity of Xpert assay with stool and pulmonary samples were 50 (95% confidence interval [CI]: 18.71–81.29%) and 100% (95% CI: 94.48–100%), respectively. Conclusion The Xpert assay on stool samples showed limited sensitivity and good specificity in the diagnosis of pulmonary TB. Therefore, it can be proposed as an alternative screening sample to diagnose TB in pediatric cases for which getting a respiratory sample is extremely difficult. However, further studies with greater number of samples and multiple baseline variables are required to support our findings. Strategies to optimize stool Xpert assay should be performed to enhance the sensitivity of this method to detect Mycobacterium tuberculosis in children.