Appropriate nodularity in ductile iron castings is strongly associated with the presence of high enough not combined Mg dissolved in the melt to cast. However, the residual Mg which is commonly measured for production control accounts for both dissolved Mg and Mg combined as oxides and sulfides. To account for the uncertainties associated with such a control, it is quite usual to over treat the melt with the risk of porosity appearance. A new methodology based on thermal analysis has been developed in the present work so as to estimate the amount of free Mg dissolved in the melt ready for pouring. A combination of Te mixture and a new “reactive mixture” composed of sulfur plus a commercial inoculant has been prepared for this purpose. This reactive mixture is able to transform the magnesium remaining dissolved in the melt to combined forms of this element. Experiments performed both during start of production (when Mg overtreatment is usual) and during normal mass production indicate that important variations of free Mg occur without relevant changes in residual Mg content as determined by spectrometry. The method developed in the present work has shown to be highly effective to detect those melt batches where active Mg content is not high enough for guaranteeing a correct nodularity of castings. Selection of proper active Mg thresholds and a correct inoculation process are critical to avoid “false”-negative results when using this new method.