This paper aims to model tourism attractiveness for the 20 Italian regions by using interregional tourist flows. This comparing some modelling approach often used when origin-destination flows data are available. To this purpose we consider the Italian domestic tourism flows and a wide range of determinants within the theoretical framework referring to the destination competitiveness theories. Using the same set of covariates (selected from Istat and Enac), we propose a comparison between the Gravity model (commonly used in tourism research) and the Bradley Terry modelling approach (to date not yet used for tourism). In particular, for the Gravity model the bilateral flows (between two regions) are directly proportional to the regions’ economic masses and inversely proportional to the distance between them, whilst in the Bradley-Terry framework, the regional tourism attractiveness is an unobservable variable that has to be scaled in a continuum on the basis of the results of multiple assessments between pairs of competitor regions. Furthermore, if the Gravity model take into account also the internal flows (on the same region), the Bradley Terry model do not consider them to model attractiveness. Consequently, different model specifications are needed and different empirical findings are obtained. Strengths and weaknesses of both modelling approaches will be analysed and explained.