The study deals with the forecast of tunnel drainage processes in discontinuous rock masses and, in particular, it is aimed to quantify the influence that some geo-structural parameters (i.e. discontinuities dip and dip direction, aperture and spacing), hydrogeological features (recharge, water table altitude, piezometric gradient) and tunnel characteristics (depth and radius) have on the tunnel drainage processes. At this aim a discreet network flow modeling (UDEC 2D) was carried out with a parametrical approach, in order to quantify the tunnel inflow and the water table drawdown with specific reference to the case of fractured and anisotropic rock masses. The tunnel inflows was also calculated using analytic formula, valid for infinite, homogeneous and isotropic aquifer, in which the permeability value is given as a modulus of equivalent hydraulic conductivity Keq. Therefore, it was observed that the numerical modeling results and the tunnel inflow calculated by analytic equations differ by over one order of magnitude. More in detail the following aspects were pointed out: the geological-structural setting critical for hydrogeological risk in tunnel, the influence of rock mass anisotropy on tunnel drainage processes, and the reliability of analytic formulas for the tunnel inflow assessment in discontinuous rock masses. On the basis of such considerations a method to integrate the analytic equations with the numerical modeling results is presented, in order to adapt the first ones to the specific hydrogeological and geo-structural condition. With regard to that, the numerical simulations were aimed to create a sufficient data set of tunnel inflows, in different geological-structural setting, enabling a quantitative comparison between numerical and analytic evaluations. In this way a correction of the traditional analytic equation of Goodman was pointed out, setting-up an empirical formula in which tunnel inflow explicitly depends on the geostructural setting of the rock mass. Finally, the obtained empirical equation was applied in a case study of a medium depth tunnel of5.5 km length, excavated in sedimentary rocks without waterproofing. The tunnel inflow was calculated both with the traditional equations and using the above cited empirical formula. The results, compared with the available monitoring data, showed that the traditional analytic equations give the highest overestimation for the stretches in which the hydraulic conductivity shows great anisotropy. On the contrary, the corrected empirical relation allows an estimation of the tunnel inflow that better reproduces the observed values.