Establishing the test–retest reliability of a motion capture system is important to support repeated walking balance assessments and to discern the relevance of significant changes to measured outcomes. Reliability of walking outcomes could also be impacted by the number of consecutive strides analysed. The primary objective of this study was to evaluate the test–retest reliability of the XSens Awinda inertial sensor-based system for total body angular momentum range, margin of stability, and step width. Twenty-eight young adults (24 ± 4 years) completed two data collections > 48 hours apart, consisting of 10 walking trials in a laboratory (8 meters) and a hallway (20 meters). Total body kinematic data from 17 sensors were used to calculate ranges of total body angular momentum (H), medial–lateral margin of stability (MOS_ML), and step width (SW). Intra-class correlations (ICC ± 95% confidence intervals) and minimum detectable change values at a 95% confidence level (MDC₉₅) were calculated using different stride counts. ICCs indicated moderate to excellent reliability for ranges of H, MOS_ML, and SW. MDC₉₅ levels were small for ranges of H and large for MOS_ML and SW. ICCs were greater in the laboratory for MOS_ML and similar between the laboratory and hallway for ranges of H and SW. MDC₉₅ indicated better reliability for the hallway suggesting that settings with a longer walkway may be ideal for reliable outcomes. The greatest improvements to reliability occurred within the first 10 consecutive strides, indicating at least 10 consecutive strides across multiple trials are recommended for improved reliability.