In this study, we assessed the possibility of using the GREENLAB model to quantitatively simulate adaptive responses of tomato plants competing for light. The model was able to reproduce modifications induced by a reduction of plant spacing in dry matter production and in the allocation pattern. However, the analysis of parameter stability showed that some parameters were environment dependent and that an adjustment of model formalisms was required to explicitly represent the plant response. Model parameters also proved to be sensitive to the data acquisition procedure. We confirmed that target files composed of dry matter records lead to parameters that were better indicators of the carbohydrates metabolism at all densities.