Abstract Context Most of the effort dedicated to the conservation of biodiversity in the European Union is applied through the establishment and maintenance of the Natura 2000 network, the world’s most extensive network of conservation areas. European Member State must actively manage these sites and report the state of the species listed in the Annexes of the Habitat and Birds Directives. Fulfilling these duties is a challenging task, especially when money available for conservation is limited. Consequently, how to optimise the use of the available economic resources is a primary goal for reserve managers. Aims In the present study, we focussed on data-sharing, and we analysed whether data-sharing among institutions may boost the performance of habitat suitability models (HSMs). Methods We collected presence data about three species of reptiles in three different protected areas of northern Italy. Then, we built HSMs under the following two different data-sharing policies: data-sharing of species’ occurrence among the different managers of the protected areas, and not sharing the occurrence data among the different managers. To evaluate how sharing the occurrence data influences the reliability of HSMs in various situations, we compared model performances under several sampling-effort levels. Key results Results show that data-sharing is usually the best strategy. In most cases, models built under the data-sharing (DS) strategy showed better performance than did data-un-sharing (DU) models. The data-sharing strategy showed advantages in model performance, notably at low levels of sampling effort. Conclusions Overcoming administrative barriers and share data among different managers of protected areas allows obtaining more biologically meaningful results. Implications Data-sharing among protected areas could allow improving the reliability of future management actions within the Natura 2000 network.