In cardiomyocytes, efficient Ca2+ release from ryanodine receptors (RyRs) is essential for the triggering of rapid and forceful contraction. The elemental units of Ca2+ release, called Ca2+ sparks, result from the concerted opening of RyRs within functional groupings called Calcium Release Units (CRUs). We have recently shown that the organization of RyRs within the CRU is altered during heart failure (HF), as RyR dispersion results in smaller and more fragmented CRUs (Kolstad et al., 2018). This reorganization was linked to slowing of both spontaneous and triggered Ca2+ sparks, and increased diastolic Ca2+ leak. Since previous work has also implicated a role of Ca2+/calmodulin-dependent protein kinase II (CaMKII) in dysregulated RyR activity during HF, we presently examined whether CaMKII activation is linked to functional CRU reorganization in this disease. Ventricular myocytes were isolated from rats which had developed HF 6 weeks following myocardial infarction. Using a newly-developed method for quantitative 3D dSTORM microscopy (Shen et al., 2019), we observed that internal RyR clusters in failing myocytes were smaller than those in cells from sham-operated animals (5.0±0.3 vs 7.9±0.5 RyRs/cluster, P<0.05). CRUs, defined as RyR clusters with edge-to-edge distances <100 nm, also displayed more distributed 3D conformations in HF, with a higher density of clusters but fewer RyRs overall. Western blotting indicated augmented RyR phosphorylation by CaMKII at serine-2814 in failing hearts. Treatment of isolated cells with the CaMKII inhibitor auto-inhibitory peptide (AIP, 10 µM) significantly reversed the reduction in RyR cluster and CRU sizes in failing cells (RyRs/CRU = 20.8±1.1, 15.1±1.3, 19.3±3.9 in sham, HF, and HF+AIP). Although AIP treatment did not alter RyR configuration in sham cells at baseline, RyR dispersion could be induced in non-failing cells by prolonged exposure to 100 nM isoproterenol. These effects which were mitigated in the presence of AIP, are consistent with a direct role of CaMKII in controlling RyR organization. Live-cell imaging of failing myocytes revealed Ca2+ sparks with slowed kinetics, as Ca2+ propagated through dispersed CRUs (time to peak =7.3±0.8 vs 10.5±0.4 ms in Sham, HF), and increased diastolic Ca2+ leak. AIP treatment accelerated spark kinetics and reduced leak to sham values. Mathematical modelling of dSTORM-derived CRU geometries supported that combined CRU fragmentation and RyR sensitization following CaMKII phosphorylation reproduced the slowed spark kinetics and increased Ca2+ leak observed experimentally in HF. Taken together, our data indicate that CaMKII-dependent RyR phosphorylation during HF is linked to a physical reorganization of the channels, which critically impairs Ca2+ homeostasis in this condition.