Background: The international advanced trauma life support guidelines recommend that all severely injured trauma patients receive supplemental oxygen based on very limited evidence. The TRAUMOX2 trial randomises adult trauma patients to a restrictive or liberal oxygen strategy for 8 h. The primary composite outcome consists of 30-day mortality and/or development of major respiratory complications (pneumonia and/or acute respiratory distress syndrome). This manuscript presents the statistical analysis plan for TRAUMOX2. Methods: Patients are randomised 1:1 in variable block sizes of four, six and eight, stratified by including centre (pre-hospital base or trauma centre) and tracheal intubation at inclusion. The trial will include 1420 patients to be able to detect a 33% relative risk reduction with the restrictive oxygen strategy of the composite primary outcome with 80% power at the 5% significance level. We will conduct modified intention-to-treat analyses on all randomised patients and per-protocol analyses for the primary composite outcome and key secondary outcomes. The primary composite outcome and two key secondary outcomes will be compared between the two allocated groups using logistic regression reported as odds ratios with 95% confidence intervals adjusted for the stratification variables as in the primary analysis. A p-value below 5% will be considered statistically significant. A Data Monitoring and Safety Committee has been established to conduct interim analyses after inclusion of 25% and 50% of the patients. Conclusion: This statistical analysis plan of the TRAUMOX2 trial will minimise bias and add transparency to the statistics applied in the analysis of the trial. The results will add evidence on restrictive and liberal supplemental oxygen strategies for trauma patients. Trial Registration: EudraCT number: 2021–000556-19; ClinicalTrials.gov identifier: NCT05146700 (date of registration: 7 December 2021).