Providing information on the impacts of climate change on hydrological processes is becoming ever more critical. Modelling and evaluating the expected changes of the water resources over different spatial and time scales can be useful in several fields, e.g. agriculture, forestry and water management. Previously a Budyko-type spatially distributed long-term climate-runoff model was developed for Hungary. This research includes the validation of the model using historical precipitation and streamflow measurements for three nested sub-catchments of the Zala River Basin (Hungary), an essential runoff contributing region to Lake Balaton (the largest shallow lake in Central Europe). The differences between the calculated (from water balance) and the estimated (by the model) mean annual evapotranspiration varied between 0.4% and 3.6% in the validation periods in the sub-catchments examined. Predictions of the main components of the water balance (evapotranspiration and runoff) for the Zala Basin are also presented in this study using precipitation and temperature results of 12 regional climate model simulations (A1B scenario) as input data. According to the projections, the mean annual temperature will be higher from period to period (2011–2040, 2041–2070, 2071–2100), while the change of the annual precipitation sum is not significant. The mean annual evapotranspiration rate is expected to increase slightly during the 21st century, while for runoff a substantial decrease can be anticipated which may exceed 40% by 2071–2100 relative to the reference period (1981–2010). As a result of this predicted reduction, the runoff from the Zala Basin may not be enough to balance the increased evaporation rate of Lake Balaton, transforming it into a closed lake without outflow.