The strategic role of logistics is increasingly being recognised, and inventory strategy is an essential element of logistics strategy. Demand forecasting is one of the most crucial issues of inventory management as forecasts form the basis for the planning of production, transportation and inventory levels. The traditional methods used for forecasting include time series methods and causal methods. In the paper, we present an Analytic Hierarchy Process-based approach to demand forecasting. The proposed decision support system offers many improvements compared to traditional methods, such as the possibility to include both tangible and intangible factors in the forecasting process, and the ability to make predictions about the future development of the environmental factors.