Abstract: To effectively reduce carbon emissions during virtual power plant operations and address the scheduling challenges posed by renewable energy volatility, this paper proposes a two-stage robust optimization model based on a tiered carbon trading mechanism, aiming to balance low carbon emissions, robustness, and economic efficiency. First, a virtual power plant system model integrating power-to-gas technology and carbon capture equipment is constructed, enhancing carbon reduction flexibility by decoupling CO₂ capture and processing stages. Second, an adjustable uncertainty set models renewable energy volatility, enabling a robust optimization framework with objectives to minimize energy procurement, carbon, and demand response costs under worst-case scenarios. This framework is solved by using a column-and-constraint generation algorithm (C&CG) to ensure optimal scheduling under uncertainty. Finally, a tiered carbon trading mechanism refines power dispatch constraints to mitigate economic losses from overly conservative scheduling. Case studies demonstrate that this model enhances system resilience and reduces economic losses while achieving a balanced low-carbon, economically efficient operational objective through the tiered carbon trading mechanism.