Objective. To determine the optimal frequency and site of stimulation for transcutaneous vagus nerve stimulation (tVNS) to induce acute changes in the autonomic profile (heart rate (HR), heart rate variability (HRV)) in healthy subjects (HS) and patients with heart failure (HF). Approach. We designed three single-blind, randomized, cross-over studies: (1) to compare the acute effect of left tVNS at 25 Hz and 10 Hz (n = 29, age 60 ± 7 years), (2) to compare the acute effect of left and right tVNS at the best frequency identified in study 1 (n = 28 age 61 ± 7 years), and (3) to compare the acute effect of the identified optimal stimulation protocol with sham stimulation in HS and HF patients (n = 30, age 59 ± 5 years, and n = 32, age 63 ± 7 years, respectively). Main results. In study 1, left tragus stimulation at 25 Hz was more effective than stimulation at 10 Hz in decreasing HR (−1.0 ± 1.2 bpm, p < 0.001 and −0.5 ± 1.6 bpm, respectively) and inducing vagal effects (significant increase in RMSSD, and HF power). In study 2, the HR reduction was greater with left than right tragus stimulation (−0.9 ± 1.5 bpm, p < 0.01 and −0.3 ± 1.4 bpm, respectively). In study 3 in HS, left tVNS at 25 Hz significantly reduced HR, whereas sham stimulation did not (−1.1 ± 1.2 bpm, p< 0.01 and −0.2 ± 2.9 bpm, respectively). In HF patients, both active and sham stimulation produced negligible effects. Significance. Left tVNS at 25 Hz is effective in acute modulation of cardiovascular autonomic control (HR, HRV) in HS but not in HF patients (NCT05789147).