We assessed whether muscle metabolism biomarkers (MMb) identified by near-infrared spectroscopy (NIRS) are valid for determining adaptations following revascularization or exercise training in peripheral artery disease (PAD). Eighteen patients (males n = 13; 69 ± 7 years) were randomized to receive revascularization (Rev = 6) or pain-free home-based exercise (Ex = 12). MMb were safely collected via a NIRS-assisted treadmill test as area-under-curve for the spectra of oxygenated (-oxy), deoxygenated (-deoxy), differential (-diff) and total (-tot) hemoglobin traces. MMb, ankle–brachial index (ABI), pain-free (PFWD) and 6-min (6MWD) walking distances were assessed at baseline and after four months. MMb were correlated at baseline with ABI (MMb-oxy r = 0.46) and 6MWD (MMb-tot r = 0.51). After treatments, MMb-oxy showed an expected increase, which was more relevant for Rev group than the Ex (56% vs. 20%), with trends towards normalization for the other MMb. These changes were significantly correlated with variations in ABI (MMb-oxy r = 0.71; p = 0.002) and 6MWD (MMb-tot r = 0.58; p = 0.003). The MMb-diff in Rev group and MMb-deoxy in Ex group at baseline predicted clinical outcomes being correlated with PFWD improvements after 4-month (r = −0.94; p = 0.005 and r = −0.57; p = 0.05, respectively). A noninvasive NIRS-based test, feasible in a clinical setting, identified muscle metabolism biomarkers in PAD. The novel MMb were associated with validated outcome measures, selectively modified after different interventions and able to predict long-term functional improvements after surgery or exercise training.