Radon and its short progeny (218Po, 214Pb, 214Bi and 214Po) are important radioactive indoor air pollutants that are well recognised for their impact on humans. Thermal spas are indoor environments which are identified as significant sources of human radiation burden due to bathing and working. Especially the transient radon and progeny concentration peaks have gained scientific attention because these were associated with short-term impact in patients and personnel. Between 2007 and 2013, novel first-order modelling was achieved for the transient concentration peaks of both radon and progeny. This type of modelling is based on a dynamical set of first-order differential equations describing radon generation and decay. These equations are combined with measurements and several reference values constituting a so-called semi-empirical approach. Real-data are used as model inputs. These are utilised through numerical modelling and the use of the Levenberg-Marquard method in estimating progeny concentrations in non-measured time moments. Through these, several exposure and dosimetric quantities are calculated. In this work, the model will be presented along with several verifications derived from the spas of Ikaria, Loutraki, and Lesvos Island (Greece). It worth to note that through this type of modelling several non-easily measured parameters -such as the attachment rate and deposition rate constants- are retrieved. Attachment rate constants ranged between 0.2 and 55 h-1. Deposition rate constants were found different between the short-lived radon decay, namely between 0.2 and 8 h-1 for attached nuclei and 0.4 and 64 h-1 for unattached nuclei. Mean annual effective doses range between 0.001 mSv and 0.6 mSv for patients and 0.001 mSv and 19 mSv for personnel. Differentiations were observed between the various spa centers depending on the radon potential of the thermal waters, the water entrance techniques, the building characteristics and other parameters.