The aim of this paper is to develop a methodology for assessing the value of water in the different stages in the water cycle. It is hypothesised that if a cubic metre of water provides some benefit in some spot at a certain moment, this cubic metre of water has a certain value not only at that point in space and time, but in its previous stages within the water cycle as well. This means that, while water particles flow from upstream to downstream, water values ‘flow’ in exactly the opposite direction. The value of water in a certain place is equal to its value in situ plus an accumulated value derived from downstream. This value-flow concept is elaborated for the Zambezi basin. It is found that water produces the smallest direct economic benefits in the upper part of the Zambezi basin. However, water flows in this part of the basin − due to their upstream location − have the highest indirect values. Return flows from the water-using sectors are particularly valuable in the upstream sub-basins. The analysis shows that the value per unit of river water increases if we go from downstream to upstream. Another finding of the study is that percolation of rainwater is generally more valuable than surface runoff. Finally, a plan to export water from the river Zambezi to South Africa is evaluated in terms of its opportunity costs. The results of this study show that the value-flow concept offers the possibility of accounting for the cyclic nature of water when estimating its value. It is stressed, however, that for the current study many crude assumptions had to be made, so that the exact numbers presented should be regarded with extreme caution. Further research is necessary to provide more precise and validated estimates.