Abstract: Decarbonizing freight transportation is a global imperative. Hydrogen fuel cell trucks (HFCTs) have emerged as a promising solution to reduce greenhouse gas emissions for heavy-duty trucking. However, widespread HFCT adoption is currently hindered by the high initial purchase cost and underdeveloped hydrogen refueling infrastructure. This study develops a dynamic bi-level Stackelberg game (DBSG) model to investigate the optimal government subsidy to support HFCT adoption in a multi-stakeholder setting. In the DBSG model, the upper-level represents the government whose objective is maximizing the $CO_2$ emission reduction when replacing traditional diesel trucks with HFCTs through subsidy allocation, while the lower-level captures the strategic interactions between two markets, HFCT manufacturing market and hydrogen refueling market, in which players are seeking to maximize their profit given the subsidy policy. By integrating both HFCT adoption demand model and hydrogen refueling demand model, the DBSG model determines the optimal subsidy amounts dynamically over the course of the adoption horizon. We formulate the DBSG as a nonlinear mathematical program. By expressing the lower-level equilibrium conditions as a system of variational inequalities, the bi-level problem is reformulated as a mathematical program with complementarity constraints (MPCC). To address the inherent degeneracy of MPCCs, a continuation-based Scholtes approximation algorithm is employed to ensure numerical tractability. Numerical results provide useful insights for policymakers to design effective incentive schemes that balance public investment and industry viability.