Background: The relationship between resting cardiac indices and exercise capacity in older adults was still not well understood. New developments in cardiac magnetic resonance imaging (MRI) enable a much fuller assessment of cardiac characteristics. Purpose/Hypothesis: To assess the association between exercise capacity and specific aspects of resting cardiac structure, function, and tissue. Study Type: Cross-sectional study. Population: A total of 112 well-functioning older adults (mean age 69 years, 52 men). Field Strength/Sequence: All participants underwent 3.0 T MRI, using scan protocols including balanced steady-state free precession cine sequence, modified look-locker inversion recovery, and T2-prepared single-shot balanced steady-state free precession. Assessment: Demographic and geriatric characteristics were collected. Blood samples were assayed for lipid and glucose related biomarkers. All participants performed a symptom-limited cardiopulmonary exercise test to achieve peakVO 2 . Cardiac MRI parameters were measured with semi-automatic software by S.Y., an 18-year experienced radiologist. Statistical Tests: Demographic, geriatric characteristics and MR measurements were compared among quartiles of peakVO2, with different methods according to the data type. Spearman's partial correlation and least absolute shrinkage selection operator regression were performed to select significant MR features associated with peakVO 2 . Mediation effect analysis was conducted to test any indirect connection between age and peakVO 2 . A two-sided P value of <0.05 was defined statistical significance. Results: Epicardial fat volume, left atrial volume indexed to height, right ventricular end-systolic volume indexed to body surface area and global circumferential strain (GCS) were correlated with peakVO 2 (regression coefficients were À0.040, À0.093, 0.127, and 0.408, respectively). Mediation analysis showed that the total effect of peakVO 2 change was 43.6% from the change of age. The proportion of indirect effect from epicardial fat volume and GCS were 11.8% and 15.1% in total effect, respectively. Data Conclusion: PeakVO 2 was associated with epicardial fat volume, left atrial volume, right ventricular volume and GCS of left ventricle.