Impaired cardiorespiratory fitness (CRF) in heart failure (HF) is influenced by a complex array of cardiac and extracardiac factors. The study aimed to identify clinical determinants of CRF measured as peak oxygen consumption (peak VO2) in HF patients, and to determine a peak VO2 prediction model using regression equations. Retrospective analysis of 200 HF patients who completed treadmill cardiopulmonary exercise testing and underwent Doppler echocardiography and/or biomarker analysis on the same day was performed. After univariate linear regression analysis, a multivariate peak VO2 prediction model was developed using significant variables in a stepwise linear regression analysis. In subjects with repeated testing, Pearson's correlation was used to assess correlations between measured and predicted change in peak VO2 (Δpeak VO2) over time. Mean age was 57 years, with 55% being male. Stepwise linear regression was used to generate a weighted model for peak VO2: 30.895 + (−0.112•age[years]) + (0.296•hemoglobin [g/dl]) + (−0.101•E/e'[unit change]) + (−0.202• body mass index [kg/m2]) + (−0.593• N-terminal pro-brain natriuretic peptide [logN pg/ml])) + (−1.349•CRP [log mg/L]). Predicted peak VO2 correlated strongly with measured peak VO2 in HF with reduced ejection fraction and HF with preserved ejection fraction patients (r = +0.63, p <0.001; r = +0.64, p <0.001, respectively). Predicted Δpeak VO2 correlated with measured Δpeak VO2 (r = +0.23, p <0.001). In conclusion, in patients with HF across a wide range of left ventricular ejection fraction, age, systemic inflammation, oxygen carrying capacity, obesity, and elevated filling pressures are the strongest predictors of impaired CRF. The proposed CRF model allows prediction of peak VO2 in HF patients and may be used to estimate peak VO2 changes over time.