Physical performance decrements observed during multi‐stressor military operations may be attributed, in part, to cellular membrane dysfunction, which is quantifiable using phase angle (PhA) derived from bioelectrical impedance analysis (BIA). Positive relationships between PhA and performance have been previously reported in cross‐sectional studies and following longitudinal exercise training programs, but whether changes in PhA are indicative of acute decrements in performance during military operations is unknown. Data from the Optimizing Performance for Soldiers II study, a clinical trial examining the effects of exogenous testosterone administration on body composition and performance during military stress, was used to evaluate changes in PhA and their associations with physical performance. Recreationally active, healthy males (n = 34; 26.6 ± 4.3 years; 77.9 ± 12.4 kg) were randomized to receive testosterone undecanoate or placebo before a 20‐day simulated military operation, which was followed by a 23‐day recovery period. PhA of the whole‐body (Whole) and legs (Legs) and physical performance were measured before (PRE) and after (POST) the simulated military operation as well as in recovery (REC). Independent of treatment, PhAWhole and PhALegs decreased from PRE to POST (p < 0.001), and PhALegs, but not PhAWhole, remained lower at REC than PRE. PhAWhole at PRE and REC were associated with vertical jump height and Wingate peak power (p < 0.001–0.050), and PhAWhole at PRE was also associated with 3‐RM deadlift mass (p = 0.006). However, PhA at POST and changes in PhA from PRE to POST were not correlated with any performance measure (p > 0.05). Additionally, PhA was not associated with aerobic performance at any timepoint. In conclusion, reduced PhA from PRE to POST provides indirect evidence of cellular membrane disruption. Associations between PhA and strength and power were only evident at PRE and REC, suggesting PhA may be a useful indicator of strength and power, but not aerobic capacity, in non‐stressed conditions, and not a reliable indicator of physical performance during severe physiological stress.