This paper shows the results of a scientific research in which a GNSS continuous monitoring system for earth-dam deformations has been developed, then, deformations have been related with reservoir water surface and level. The experiment was conducted near Bivona (Sicily, Italy), on the Castello dam (Magazzolo Lake). On the top of the dam three control points were placed and three GNSS permanent stations were installed. The three stations continuously transmitted data to the control centre of the University of Palermo. The former has been determined using freely available satellite data (specifically Landsat 7 SLC-Off) collected during the whole study period (DOYs 101 to 348 2011). Issues related with the un-scanned rows filling and to better distinguish water from land pixels on the shoreline. The aim of this work is various: first of all, we want to evaluate whether the GPS post processing techniques can provide static results comparable to other monitoring techniques, such as spirit levelling. The study could take a significant importance given that the Italian legislation until today does not provide for the use of this technology to manage or monitor dams displacements or other civil engineering constructions. The use of GPS data in structural monitoring could in fact reduce some management costs. Usually the conventional GPS monitoring methods, where a base station GPS receiver must be located near the dam, did not ensure that the accuracy of results have been independent from the displacement of the crown (top end of dam). In this paper, a new approach in the area of study of the GNSS permanent network has been engaged to solve these problems. Field-testing results show that the new GNSS approach has excellent performances, and the monitoring of different section of the dam could reveal important information on its deformation, that its not operationally possible to retrieve elsewhere. The post-processing accuracy positioning is around 1–5 mm for the deformations monitoring of the Castello dam. Displacements of different sections of the dam reveal different behaviour (in time and periodicity) that looks to be related with water surface (and level) retrieved from remote sensing.