In heart failure (HF) with sinus rhythm, resting and exercise heart rates correlate with exercise capacity and mortality. However, in HF with atrial fibrillation (AF), this correlation is unknown. Our aim is to investigate the association of resting and exercise ventricular rates (VRs) with exercise capacity and mortality in HF with AF. We identified 903 patients with HF and AF referred for cardiopulmonary stress testing. AF was defined as history of AF and AF during cardiopulmonary stress testing. We constructed multivariable models to evaluate the association of resting VR, peak exercise VR, VR reserve (peak VR-resting VR), and chronotropic index with (1) peak oxygen consumption (PVO2) ≤18 mL/kg per minute, (2) continuous PVO2, and (3) 10-year all-cause mortality. Median (25th-75th percentile) age was 60 (52-67) years, left ventricular ejection fraction was 25 (15-50)%, and 76.1% were males. Patients with lower (quartile 1) compared with higher (quartile 4) peak VR, VR reserve, and chronotropic index were more likely to have PVO2 ≤18 mL/kg per min (adjusted odds ratio [95% CI]: 14.92 [8.07-27.58], 24.60 [12.36-48.98], and 22.31 [11.24-44.27], respectively), and higher all-cause mortality (adjusted hazard ratio [95% CI]: 2.56 [1.62-4.04], 2.29 [1.47-3.59], and 2.30 [1.51-3.49], respectively). For every 10 beats per minute increase in VR reserve, PVO2 increased by 1.05 mL/kg per minute (B-coefficient [95% CI]: 1.05 [0.94-1.15]) and mortality decreased by 12% (adjusted hazard ratio [95% CI]: 0.88 [0.83-0.94]). Resting VR was associated with PVO2 (B-coefficient [95% CI]: -0.46 [-0.70 to -0.23]) but not mortality (adjusted hazard ratio [95% CI]: 0.97 [0.88-1.06]). In patients with HF and AF, higher resting VR and lower peak exercise VR, VR reserve, and chronotropic index were all associated with worse peak exercise capacity, but only lower exercise VR parameters were associated with higher mortality. Dedicated studies are needed to gauge whether modulating exercise VR enhances exercise performance and outcomes.