Reducing carbon emissions and increasing the integration of new energy sources are key steps towards achieving sustainable development. Virtual power plants (VPPs) play a significant role in enhancing grid security and promoting the transition to clean, low-carbon energy. The core equipment of the VPP, the CHP unit, utilizes a thermal engine or power station to generate electricity and useful heat simultaneously. However, the intermittent and volatile nature of renewable energy sources, as well as the “heat-driven power generation” mode of combined heat and power (CHP) units, presents contradictions that severely affect their peak-shifting capability and lead to high carbon emissions. To address these issues, a novel VPP is established by integrating traditional power plants with carbon capture and hydrogen energy storage. This approach utilizes a “hydrogen energy storage−electric boiler” decoupling method to address the operational mode of CHP, strengthens the coupling relationship between electric and thermal hydrogen loads, and considers a tiered carbon-trading mechanism. With the net profit of the VPP as the optimization objective, the model balances economic and environmental considerations and establishes a low-carbon economic dispatch model for the VPP. A genetic algorithm is employed for solving, and three different dispatch strategies are set for simulation in three distinct seasonal scenarios. The comprehensive comparative analysis of the dispatch results reveals a reduction in carbon emissions and an increase in net profit to varying degrees across all three seasons. Overall, the proposed dispatch strategy demonstrates the ability to enhance the new energy-integration capacity and total revenue of a VPP while simultaneously achieving the goal of reducing carbon emissions.