The protection of people, buildings and infrastructure against natural hazards is one of the key functions of mountain forests. Since the protective function strongly depends on small-scale local conditions such as terrain and stand characteristics, spatially explicit evaluation methods are necessary to provide information required for an effective and cost-efficient forest management. Risk analyses are recognized as the best method for estimating the danger from various natural hazards. Currently, however, risk-based strategies are rarely addressed in the management of protection forest. We present and discuss a risk-based approach to evaluate the protective effect of mountain forests in a spatially explicit manner to demonstrate the advantages of future risk-based protection forest management. We illustrate the approach by performing a GIS-based risk analysis in the case study area 'Bannwald of Andermatt' (Switzerland) for a 300-year snow avalanche event. Classifying forest structures based on aerial photographs allowed developing different forest cover scenarios and modeling potential avalanche release areas within the forest. Potential avalanche release areas above the forest and the avalanche run-out distances under five different scenarios of forest cover were calculated by using a two-dimensional avalanche simulation model. We calculated the annual collective risk for each scenario and compared the change in risk to reveal the spatial distribution of the protective effect of the forest. Resulting risks differed strongly between forest cover change scenarios. An enlargement of an existing wind-disturbed area within lower parts of the slope resulted only in a slight increase of risk. In contrast, the effect of an unforested area in the upper parts of the observed forest more than doubled the risk. These results show how a risk-based approach can help to quantify and illustrate the impact of differences in forest cover on the protective effect of mountain forests. It is a promising approach to determine the economic value of protection forests and thus provide quantitative and qualitative information for cost-efficient forest maintenance planning. With regard to the achievements of research to date, the presented approach may serve as a valuable method to support decision-making in a future protection forest management.