Background: Magnesium (Mg) is often used to manage de novo atrial fibrillation (AF) in the emergency department (ED) and intensive care unit (ICU). Point of care measurement of ionized magnesium (iMg) allows a rapid identification of patients with impaired magnesium status, however, unlike ionized calcium, the interpretation of iMg is not entirely understood. Thus, we evaluated iMg reference values, correlation between iMg and plasmatic magnesium (pMg), and the impact of pH and albumin variations on iMg levels. Secondary objectives were to assess the incidence of hypomagnesemia in de novo AF. Methods: A total of 236 emergency department and intensive care unit patients with de novo AF, and 198 control patients were included. Reference values were determined in the control population. Correlation and concordance between iMg and pMg were studied using calcium (ionized and plasmatic) as a control in the whole study population. The impact of albumin and pH was assessed in the discordant iMg and pMg values. Lastly, we assessed the incidence of ionized hypomagnesemia (hypoMg) among de novo AF. Results: The reference range values established in our study for iMg were: 0.48–0.65 mmol/L (the manufacturers were: 0.45–0.60 mmol/L). A strong correlation was observed between pMg and iMg (r = 0.85), but, unlike for calcium values, there was no significant impact of pH and albumin in iMg/pMg interpretation. The incidence of hypoMg among de novo AF patients was 8.5% (12.7% using our ranges). When using our ranges, we found a significant link (p = 0.01) between hyopMg and hypokalemia. Conclusion: We highlight the need for more accurate reference range values of iMg. Furthermore, our results suggest that blood Mg content is not identical to that of calcium. The incidence of ionized hypomagnesemia among de novo AF patients in our study is 8.5%.