Accelerometer threshold values to detect physical activity intensity in the oldest old persons with coronary artery disease (CAD) are lacking as well as knowledge about their free-living physical activity behavior. The purpose of this study was 2-fold. (1) To assess the sensitivity and specificity of 3 existing intensity threshold values for ActiGraph accelerometers for the oldest old persons with CAD. (2) To assess free-living physical activity, applying the threshold values with the highest sensitivity and specificity for assessing at least moderate intensity, among these persons. In a cross-sectional design, a total of 24 persons with CAD, mean age 87.5 (3.7) years, participated in the study at a university hospital in Sweden. To assess the sensitivity and specificity of the existing threshold values, the participants walked at different speeds wearing the accelerometer at a pace corresponding to individualized perceived exertion at light, moderate, and high intensity according to the Borg Rating of Perceived Exertion Scale. For the free-living physical activity assessment, the persons wore the accelerometer for 7 consecutive days. The percentage agreement for light-, moderate-, and high-intensity threshold values, as well as receiver operating characteristic curves, was used to identify the sensitivity and specificity of the existing threshold values for moderate intensity. The threshold values for at least moderate intensity at 1041 counts per minute according to Copeland had the highest sensitivity (0.739) and specificity (0.609) to identify at least moderate intensity for the ActiGraph GT3X+ accelerometer. In a free-living setting, the oldest old persons with CAD spent 11 of 13.5 (81%) waking hours in a sedentary position and, of the 2.5 hours of being active, 19 minutes (2%) were at least at moderate intensity. Nine of 24 persons (38%) reached 20 minutes of moderate- to vigorous-intensity physical activity 3 days a week, according to guidelines for exercise-based cardiac rehabilitation. The existing threshold values according to Copeland had the highest sensitivity and specificity to identify at least moderate intensity and are valid for use in the oldest old persons with CAD. Using accelerometry as an objective measurement for physical activity can help further improve our understanding of free-living physical activity behavior and to assess relationships between free-living physical activity and health outcomes among the oldest old persons with CAD.