Simulation models of the processes that control carbohydrate balance in coniferous trees are reviewed, and their appropriateness for assessing pollution effects is considered. Currently such models are at the forefront of attempts to simulate the growth process of trees, but they are not able to predict growth of forest trees accurately under varied environmental conditions such as an increased pollution load. Typically model structures for growth based on carbon balance incorporate such features as constant allocation coefficients or a regulation of foliage amount around a theoretical maximum at canopy closure. It is these features that render such models unable to predict the effects of pollution because the processes of compensation in relation to pollution load are obscured. The study presents a model, CARBON, that prioritizes allocation between meristems according to functional requirements, but this is still insufficient to predict the effects of pollution. The study discusses the problem and that of using a maximum foliage amount as a control parameter in canopy development, and propose suggestions for an improved modeling framework for tree growth models. This should include the expression of translocation, storage and utilization of carbohydrates in a spatially well defined system.