The aim of this study was to assess the possible association between P-Mg and subsequent high levels of HbA1c. The study involves data from primary health care patients and data from patients treated in hospitals located in the capital region of Denmark. P-Mg and HbA1c levels were analyzed from 121,575 patients in the period 2010-2022. Patients were categorized in a diabetic and non-diabetic group. Out of 121,575 patients, 8,532 were categorized as diabetic. A reverse J-shaped association between P-Mg and HbA1c levels ≥ 48 mmol/mol was observed with nadir at P-Mg of 0.90 mmol/L. The unadjusted hazard ratio (HR) for having a first HbA1c measurement ≥ 48 mmol/mol is 1.54 (95% Cl 1.50; 1.57) per 0.1 mmol/L decrease in P-Mg when P-Mg is lower than 0.90 mmol/L. After adjusting for age and gender, the HR remained significant at 1.45 (95% Cl 1.41; 1.48).For P-Mg levels above 0.90 mmol/L, the unadjusted HR per 0.1 mmol/L increase in P-Mg was 1.04 (95% Cl 1.02; 1.06). After adjusting for age and gender the HR remained significant at 1.06 (95% Cl 1.05; 1.08). In conclusion, this study found that patients with dysmagnesemia have a higher risk of developing diabetes even after adjusting for age and gender. Hyper- or hypomagnesemia in patients could be a biomarker for predicting the risk of developing diabetes.