Key points Intracellular [Na+] ([Na+]i) is elevated in heart failure (HF) and causes arrhythmogenic cellular [Ca2+]i loading. In HF, hyperactivity of Ca2+–calmodulin‐dependent protein kinase II (CaMKII), a key mediator of electrical and mechanical dysfunction in myocytes, causes elevated [Na+]i. We developed a computational model of mouse ventricular myocyte electrophysiology including Ca2+ and CaMKII signalling and quantitatively confirmed evidence suggesting that not only does CaMKII cause elevated [Na+]i, but this additional [Na+]i also promotes further CaMKII activation by increasing [Ca2+]i. We found that a 3–4 mm gain in [Na+]i (similar to that reported in HF) perturbs Ca2+ and membrane potential homeostasis in part via CaMKII activation. This disrupted Ca2+ homeostasis is exacerbated by CaMKII overexpression, and strongly relies upon CaMKII–Na+–Ca2+–CaMKII feedback. CaMKII inhibition in HF may be beneficial, in part by inhibiting [Na+]i loading, and thereby normalizing Ca2+ and membrane potential dynamics without disrupting systolic function. Abstract Ca2+–calmodulin‐dependent protein kinase II (CaMKII) hyperactivity in heart failure causes intracellular Na+ ([Na+]i) loading (at least in part by enhancing the late Na+ current). This [Na+]i gain promotes intracellular Ca2+ ([Ca2+]i) overload by altering the equilibrium of the Na+–Ca2+ exchanger to impair forward‐mode (Ca2+ extrusion), and favour reverse‐mode (Ca2+ influx) exchange. In turn, this Ca2+ overload would be expected to further activate CaMKII and thereby form a pathological positive feedback loop of ever‐increasing CaMKII activity, [Na+]i, and [Ca2+]i. We developed an ionic model of the mouse ventricular myocyte to interrogate this potentially arrhythmogenic positive feedback in both control conditions and when CaMKIIδC is overexpressed as in genetically engineered mice. In control conditions, simulation of increased [Na+]i causes the expected increases in [Ca2+]i, CaMKII activity, and target phosphorylation, which degenerate into unstable Ca2+ handling and electrophysiology at high [Na+]i gain. Notably, clamping CaMKII activity to basal levels ameliorates but does not completely offset this outcome, suggesting that the increase in [Ca2+]i per se plays an important role. The effect of this CaMKII–Na+–Ca2+–CaMKII feedback is more striking in CaMKIIδC overexpression, where high [Na+]i causes delayed afterdepolarizations, which can be prevented by imposing low [Na+]i, or clamping CaMKII phosphorylation of L‐type Ca2+ channels, ryanodine receptors and phospholamban to basal levels. In this setting, Na+ loading fuels a vicious loop whereby increased CaMKII activation perturbs Ca2+ and membrane potential homeostasis. High [Na+]i is also required to produce instability when CaMKII is further activated by increased Ca2+ loading due to β‐adrenergic activation. Our results support recent experimental findings of a synergistic interaction between perturbed Na+ fluxes and CaMKII, and suggest that pharmacological inhibition of intracellular Na+ loading can contribute to normalizing Ca2+ and membrane potential dynamics in heart failure.