Diabetes mellitus and atrial fibrillation (AF) are major unsolved public health problems, and recent clinical studies show that diabetes is an independent risk factor for AF. However, the mechanism(s) underlying this clinical association are unknown. Atrial myocardium from diabetic patients is marked by increased reactive oxygen species (ROS) and elevated protein O-GlcNAcylation (OGN). Calmodulin kinase II (CaMKII) is a recognized proarrhythmic signal that may be activated by OGN (OGN-CaMKII) and ROS (ox-CaMKII). We used a combination of new and established mouse models to selectively control the activity of myocardial CaMKII, ox-CaMKII, and OGN-CaMKII. We found that diabetes significantly increased AF by a ROS, CaMKII, and ox-CaMKII pathway. In contrast, mice with a knock-in mutation eliminating a recently identified CaMKII OGN site were not protected from diabetic AF, nor did OGN enhance catalytic activity of CaMKII in vitro. In contrast, diabetic knock-in mice with ryanodine receptors resistant to CaMKII, and diabetic wild type mice treated with an OGN antagonist were protected from AF induction. Our data support a concept where ox-CaMKII, and OGN converge on the ryanodine receptor causing pathological Ca2+ leak, myocardial cell membrane hyperexcitability, and excessive AF in diabetic mice. These results provide needed insights into the mechanism for increased AF in diabetes, and suggest control of excessive ROS, OGN, or CaMKII are potential therapeutic strategies for diabetic AF.