Road blocking due to thawing or heavy rains annually contributes to a considerable loss of profit in Swedish forestry. Companies have to build large stocks of sawlogs and pulplogs to secure a continuous supply during periods where the accessibility of the road network is uncertain. This storage leads to quality deterioration, which means loss in profit. One approach to reduce the losses due to blocked roads is to upgrade the road network to a standard that guarantees accessibility throughout the year. This article describes a decision support system called RoadOpt for the planning of forest road upgrading. The planning horizon is about one decade. The system uses a Geographical Information System (GIS)-based map user interface to present and analyse data and results. Two important modules are the Swedish road database, which provides detailed information about the road network, and an optimization module consisting of a mixed integer linear programming model. A case study from a major Swedish company is presented.