AIMS: Accumulating evidence suggests Gulf War illness (GWI) is characterised by autonomic nervous system dysfunction (higher heart rate [HR], lower heart rate variability [HRV]). Yoga - an ancient mind-body practice combining mindfulness, breathwork, and physical postures - is proposed to improve autonomic dysfunction yet this remains untested in GWI. We aimed to determine (i) whether HR and HRV improve among Veterans with GWI receiving either yoga or cognitive behavioural therapy (CBT) for pain; and (ii) whether baseline autonomic functioning predicts treatment-related pain outcomes across follow-up. MAIN METHODS: We present secondary analyses of 24-hour ambulatory cardiac data (mean HR, square root of the mean squared differences between successive R-R intervals [RMSSD], high frequency power [HF-HFV], and low-to-high frequency ratio [LF/HF] extracted from a 5-min window during the first hour of sleep) from our randomised controlled trial of yoga versus CBT for pain among Veterans with GWI (ClinicalTrials.govNCT02378025; N = 75). KEY FINDINGS: Veterans who received CBT tended towards higher mean HR at end-of-treatment. Better autonomic function (lower mean HR, higher RMSSD/HF-HRV) at baseline predicted greater reductions in pain across follow-up, regardless of treatment group. Better baseline autonomic function (mid-range-to-high RMSSD/HF-HRV) also predicted greater pain reductions with yoga, while worse baseline autonomic function (higher mean HR, lower RMSSD/HF-HRV) predicted greater pain reductions with CBT. SIGNIFICANCE: To our knowledge, this is the first study to suggest that among Veterans with GWI, HR may increase with CBT yet remain stable with yoga. Furthermore, HR and HRV moderated pain outcome across follow-up for yoga and CBT.