The need to reduce dependence on fossil fuels and minimize greenhouse gas emissions has led to increasing research on sustainable renewable energy source, including bioenergy. However, commercial scale of bioenergy production is severely constrained by investment risk and uncertainty. This study evaluates the impacts of policy supports on the optimal supply chain decisions adapting to feedstock yield uncertainty given different risk preferences of the biofuel sector in an integrated system. A two-stage stochastic mixed integer linear programming is developed to model a risk neutral biofuel sector that minimizes the expected cost considering feedstock cultivation along with biorefinery establishment. Alternatively, the Conditional Value-at-Risk (CVaR) minimization focuses on the financial risk associated with investment decisions under feedstock yield uncertainty. Results of a case study in west Tennessee suggest that CVaR minimization case converts more land for switchgrass cultivation compared to the expected cost minimization to lower the high costs of low yield conditions. With the availability of subsidies, expected cost and investment risk are improved for the biofuel sector. However, both the expected cost and investment risk are reduced by a higher percent for the risk averse case than the risk neutral one that minimizes expected cost.