Stool samples are alternatives to respiratory samples for bacteriological confirmation of childhood tuberculosis but require intensive laboratory processing before molecular testing to remove PCR inhibitors and debris. We aimed to develop a centrifuge-free processing method for use in resource-limited settings based on a sucrose-flotation method that showed good sensitivity for childhood tuberculosis diagnosis. In an in vitro study using Xpert MTB/RIF Ultra on stool samples spiked with defined bacterial concentrations of Mycobacterium tuberculosis (MTB), we compared different simplification parameters to the reference sucrose-flotation method. Best methods were selected based on the rate of invalid/error results and on sensitivity, compared to the reference method on stools spiked at 10 colony forming units (CFU)/g MTB. For final selection, we tested the best parameter combinations at 10 CFU/g. Out of 13 different parameter combinations, three were tested at 10 CFU/g. The best combination used 0.5 g stool, manual shaking, no filtration, 30-min sedimentation, and a 1:3.6 dilution ratio. This method gave 10% invalid/error results and a sensitivity of 70% vs 63% at 10 CFU/g and 53% vs 58% at 10 CFU/g compared to the reference method. This pre-clinical study was able to develop a centrifuge-free processing method to facilitate stool Xpert Ultra testing.Copyright © 2020 Elsevier Ltd. All rights reserved.