With changes in the competitive structure of the electric energy market and the development of distributed generation resources and demand response loads, virtual power plant bidding has become one of the key topics. This paper investigates a probabilistic decision-making model for optimizing virtual power plant bidding in the day-ahead market, considering the participation of demand response loads in intra-day markets. The proposed model includes a three-stage programming approach for the simultaneous management of energy and reserve capacity and evaluates the impact of demand response load exchanges in intra-day markets on improving the plant’s profit while considering risk tolerance. In the proposed model, the participation of the virtual power plant in both the energy market and the intra-day market is modeled in detail using an innovative method based on the stochastic dual dynamic programming (SDDP) approach. The study examines the effect of participation in the intra-day market on changing the virtual power plant’s behavior and reducing imbalances caused by production forecast errors. The results show that participation in the intra-day market can increase the virtual power plant’s profit by up to 2.5% and improve risk management by 2.7%.