Physical activity (PA) measured by accelerometry is proposed as a novel trial endpoint for heart failure (HF). However, standardised methods and associations with established markers are lacking. This study aimed to examine PA measurements and accelerometer repeatability in patients with HF and age- and sex-matched controls, and study correlations with established prognostic HF markers, body composition, and quality of life (QoL). Accelerometry was performed in 105 patients with HF with left ventricular ejection fraction (LVEF) ≤ 40% and in 46 controls. Participants also underwent dual X-ray absorptiometry, cardiopulmonary exercise testing, a six-minute walking test (6MWT), echocardiography, and NT-proBNP measurement, and completed a QoL questionnaire. Average acceleration was markedly reduced in patients with HF compared with healthy controls (16.1 ± 4.8mg vs 27.2 ± 8.5mg, p < 0.001). Healthy controls spent a median daily 56min (IQR 41-96min) in moderate-to-vigorous PA (MVPA), whereas HF patients spent only 12min (IQR 6-24) in MVPA. In HF patients, average acceleration correlated moderately with 6MWT (R = 0.41, p < 0.001) and maximal oxygen uptake (peak VO2) (R = 0.36, p < 0.001) but not with NT-proBNP, LVEF, or QoL. Patients in NYHA class II showed a higher average acceleration than patients in NYHA III (16.6 ± 4.9mg vs 14.0 ± 3.6mg, p = 0.01). Daily PA was severely reduced in patients with HF compared with healthy controls. In HF patients, we found moderate correlations of accelerometer measurements with markers of physical capacity but not with LVEF or NT-proBNP. NCT05063955. Registered 01 June 2021-retrospectively registered.