Climate change affects Central European forest ecosystems in different ways and, consequently, these changes result in different feedbacks on the provision of forest ecosystem services. Regarding the complexity and the variability in climate-forest interactions outcome, forest decision makers necessitate reliable information about changes in the forest ecosystem services for planning and adaptation purposes. However, forest productivity predictions incorporate multiple levels of uncertainty that have to be regarded to ensure building realistic expectations in forest decision-making. Besides the chosen forest simulation model, uncertainties come from the climate change data represented by a set of representative concentration pathways (RCP), within the underlying ensemble of global circulation and regional climate models (GCM-RCM), and further in the treatment of the CO2-fertilization effect. We considered the mentioned uncertainties in a framework on simulating forest growth and water services for two forest sites, a Sessile oak and a Scots pine stand in Rhineland-Palatine, Germany. The framework revealed a high variability in future forest ecosystem services. Particularly, the variability among the selected GCM-RCM models within the same Representative Concentration Pathway (RCP) was higher than the variability among different RCPs (RCP2.6 and RCP8.5 representing the low and high CO2-emission scenarios, respectively). Sessile oak productivity increased under all scenarios, whereas Scots pine growth declined in the lower end of the RCP8.5 scenario. Water services remained mostly stable at both sites. Moreover, we applied a panel data model to estimate what climate indices caused changes in the forest ecosystem services. We found that Scots pine is more sensitive to a multitude of climate indices, such as temperature changes and Sessile oak showed strong response to the CO2-fertilization. We propose applying this framework to evaluate forest management options under climate change.