Inhaled radon from groundwater used for domestic purposes is one of the sources of natural radioactivity into indoor air. Due to uranium-bearing minerals occurrences, hydrogeochemical conditions, tectonic structures, and hydraulic circuits, the radon pathway from rocks to groundwater is quite unpredictable. High radon potential from bedrocks is not always associated with high radon levels in groundwater. Besides, inhaled radon from domestic use may also increase the exposure toindoor radon levels. This innovative methodology using hydrogeochemical conditions and groundwater flow transport was used for radon predictions in the underground to ensure safe drinking water ingestion and inhalation. This innovative radon prediction methodology is based on classic hydrogeochemical analyses (Eh-pH, Piper, Schöeller and Gibb's diagrams) and multivariate statistical analyses (Principal Component Analysis and Pearson's correlation). High dissolution of major ions does not imply high radon mobilization from rocks to groundwater. The travel time was estimated to developed a flow transport of contaminated groundwater. Radiological results show that of the 25 sampled springs, five of them contained radon concentrations above the Portuguese imposed limit (Rn = 500 Bq·L), and 16 of them with values above the WHO recommended limit (Rn = 100 Bq·L). Overall, this new approach of radon prediction showed that uranium enrichment in rocks at ideal hydrochemical conditions and emanation coefficient, and shallow circuits, are responsible for radon increasing in drinking water. The proposed approach allow to predict the areas with high radon potential groundwaters, being a tool to be used by water planners and policy makers for corrective and preventive measures in shallow groundwater flows. To safeguard clean water within the predefined deadline of Sustainable Development Goals (2030) and to ensure human health in compliance with WHO guidelines for safe drinking water, should be established priority water protection policies to reduced radon in this contaminated springs (n = 16).