Key points Cardiac sympathetic afferents are considered to be essential pathways for transmission of cardiac nociception to the central nervous system during myocardial ischaemia. However, a potential contribution of the CSAR control of cardiac dysfunction in both normal and chronic heart failure (CHF) states remains unknown. We found that activation of the CSAR evokes little increase in cardiac contractility with an exaggerated peripheral vasoconstriction in the CHF state. CSAR inhibition by epicardial lidocaine decreased cardiac contractility to a greater extent in CHF rats than sham rats. Furthermore, we also found that epicardial lidocaine paradoxically decreased left ventricular end‐diastolic pressure (LVEDP) and left ventricular end‐diastolic volume (preload) in CHF rats, which was not observed in sham rats. Chronic ablation of the CSAR by epicardial application of the afferent neurotoxin, RTX, selectively lowered diastolic blood pressure CHF rats. The observation suggests that CSAR has a differential effect on cardiac function in normal and CHF states. CSAR activation in normal state causes significant increase in cardiac contractility and cardiac output. Abstract The enhanced ‘cardiac sympathetic afferent reflex’ (CSAR) critically contributes to the exaggerated global sympathetic tone in chronic heart failure (CHF). However, a potential contribution of the cardio‐cardiac reflex control of cardiac function in both normal and CHF states remains unknown. In this study, we evaluated the effects of direct activation or inhibition of the CSAR on cardiac function by pressure–volume (P–V) loop analysis in ∼12‐week sham‐operated and myocardial infarcted (MI) rats. In sham rats, acute CSAR activation by epicardial application of bradykinin (BK) increased heart rate (HR), left ventricular systolic pressure (LVSP), the maximum first derivative of left ventricular pressure (dp/dtmax), and the slope of the end‐systolic P–V relationship (ESPVR), suggesting that acute CSAR activation in the normal state enhances myocardial contractility. CSAR activation also decreased left ventricular (LV) systolic and diastolic volumes with little effect on LV end‐diastolic pressure (LVEDP) or the end‐diastolic P–V relationship (EDPVR) in sham rats. Compared to sham, CHF rats exhibit a reduced increase in the slope of the ESPVR and dp/dtmax in response to BK, indicating a poor contractile response to CSAR activation. Interestingly, BK application in CHF rats increased cardiac systolic and diastolic volumes and further increased the elevated LVEDP, neither of which was seen in sham rats. Following CSAR inhibition by epicardial lidocaine, blood pressure, HR, LVSP, dp/dt, LVEDP and ESPVR decreased in CHF rats whereas lidocaine had little effect in sham rats, indicating that the CSAR is tonically active in CHF and contributes to cardiac dysfunction. Furthermore, we found that epicardial lidocaine paradoxically decreased LV end‐diastolic volume (preload) in CHF rats, which was not observed in sham rats. The decreased preload by lidocaine in CHF rats may be due to a reduction in peripheral vascular resistance since epicardial lidocaine significantly lowered peripheral (renal) sympathetic nerve activity in CHF rats but not in sham rats. Furthermore, chronic ablation of CSAR by epicardial application of a selective afferent neurotoxin, resiniferatoxin, selectively lowered diastolic blood pressure both at daytime and night‐time with less effect on systolic blood pressure in CHF rats. Our data suggest that there is an imbalance between cardiac and peripheral responses to CSAR in CHF animals compared to sham‐operated controls.