It is argued that the forest can provide low-cost options to reduce the atmospheric CO2 concentration. However, many dimensions of the future dynamics of the forest, and its interactions with climate change are still not well understood. This paper provides new insights into how these types of uncertainties affect the optimal climate policy. We model uncertainty over several key forest parameters by using the novel state-contingent approach. Our main results show that the importance of including optimal forest controls in climate policy increases when the dynamics of the forest are uncertain. Ignoring uncertainties concerning the forest will lead to biased estimates of the social costs of carbon and be misleading when evaluating climate policies. Conversely, recognizing forest uncertainties and its potential to mitigate climate change will lead to a robust policy where the cost of uncertainty to a large extent can be avoided.