Forestry poses a challenge to computable general equilibrium (CGE) modellers working with recursively dynamic models. This is because of the lag between its inputs and output, which do not correspond to the same time period as other sectors. Inputs are applied for a number of years before a forest is ready for harvest. As a result, attempts in the past to incorporate a well-specified forestry sector in a recursive CGE model have been only partly successful. The purpose of this paper is to fill this gap by presenting a consistent analytical framework which can be used to incorporate forestry into a recursively dynamic global CGE model. A key feature of the framework is that it splits the forestry activity into three parts—planting, holding and harvesting. Planting and harvesting are done by standard production sectors and holding is done by investors, whose behaviour is already modelled in these CGE models. In addition, global forests are classified into three groups—commercial plantation forests, environmental plantation forests and native forests. All harvested forest land is made available for competition for alternative agricultural uses and will be allocated to the activity it is best suited for, given productivity differences for different activities. This framework can be used in a CGE modelling framework to support implementation of the proposed reduced emissions from deforestation and forest degradation (REDD) schemes as well as being applied to study land allocations, nationally and globally, across activities under alternative scenario assumptions. For example, the model can be used to project the effects on food production and prices of an increase in bio-fuel subsidies.