The combined use of global positioning system (GPS), levelling and geoid height information has been a key procedure in various geodetic applications and provides at the same time a very attractive evaluation scheme for the accuracy of gravimetric geoid models. The main goal of this paper is to propose a tool of adjustment of combined GPS/Levelling/Geoid networks. For this kind of adjustment, generally, two main types of unknowns are estimated; the gravimetric geoid accuracy and 2D spatial field that describe all the datum/systematic distortions between the available height data sets. Accordingly, two modelling alternatives for the correction field are programmed, namely a pure deterministic parametric model, and a hybrid deterministic and stochastic model. In this context, and for a first attempt, a Fortran program has been developed translating in application this methodology using the Hirvonen model as analytic covariance function of the reduced signals and the four parameter model to describe all possible datum inconsistencies and others systematic effects between the available height data sets. The necessary data used to test this program, the adopted methodology, the computation procedure, and the obtained results will be presented.