HYPOTHESIS: This study hypothesized that augmented reality (AR) technology has comparable accuracy and safety to conventional CT localization in guiding percutaneous transthoracic lung puncture (PTLP) to localize small pulmonary nodules. METHODS: This study was a prospective, non-inferiority randomized clinical trial. Patients were randomly assigned between 23 May 2023, and 26 September 2023. Patients with small peripheral lung nodules (≤2 cm) were recruited. Patients were randomly assigned to either the CT-guided PTLP group or the AR-guided PTLP group, with a 1:1 allocation ratio. The primary outcome was the accuracy of lung nodule localization measured by localization error. The secondary outcomes included procedure duration, radiation exposure dosage and complications. RESULTS: A total of 70 patients underwent either CT- or AR-guided lung nodule localization and subsequent surgeries. Localization error was smaller in the AR-guided group than in the CT-guided group (mean ± SD, 3.1 ± 4.0 mm vs. 5.4 ± 4.2 mm, P = 0.026). The mean difference of localization errors was -2.3 mm (95% CI: - 4.2 to -0.3 mm, P < 0.001 for non-inferiority). Compared to the CT-guided group, the AR-guided group demonstrated significantly lower radiation exposure (mean ± SD, 421 ± 168 vs. 694 ± 229 mGy × cm, P < 0.001) and shorter localization procedure duration (mean ± SD, 8.8 ± 2.3 vs. 14.1 ± 1.8 minutes, P < 0.001), with no statistical difference in complications. CONCLUSIONS: The accuracy of the AR-guided approach is comparable to that of the CT-guided approach in localizing small lung nodules. Furthermore, the utilization of AR technology has been demonstrated to reduce procedural time and minimize radiation exposure for patients.