Although physical activity (PA) is an important determinant of exercise capacity, the association between these constructs is modest. The authors investigated the associations of self-reported and objectively measured PA with maximal and submaximal tests of exercise capacity. Participants aged ≥40 years (N = 413; 49.6% female) completed a PA questionnaire, wore a uniaxial accelerometer (5.2 ± 1.1 days), and performed maximal (cardiopulmonary exercise test [CPET]) and submaximal (long-distance corridor walk) tests with indirect calorimetry (oxygen consumption, $\dot{\mathrm{V}}{\mathrm{O}}_2$). Linear regression models were fitted to assess the variation in exercise capacity explained (partial eta squared, η²) by PA variables. Accelerometer-measured vigorous (η² = 22% female; η² = 16% male) and total PA (η² = 17% female; η² = 13% male) explained the most variance in CPET $\dot{\mathrm{V}}{\mathrm{O}}_2$ (p < .001). All η² values were lower for long-distance corridor walk $\dot{\mathrm{V}}{\mathrm{O}}_2$ (η² ≤ 11%). Age contributed more to CPET $\dot{\mathrm{V}}{\mathrm{O}}_2$ than any PA variable in males (η² = 32%), but not in females (η² = 19%). Vigorous and total PA play important roles in CPET $\dot{\mathrm{V}}{\mathrm{O}}_2$ in mid to late life.