With the rapid development of new power systems, virtual power plants (VPPs) have become a significant technical approach for integrating large-scale distributed energy resources into electricity market transactions. However, the uncertainties in renewable energy output and electricity market prices increase the coupling complexity of decision spaces among various stakeholders within a VPP, posing significant challenges to its optimal operation. To address this, a multi-stakeholder collaborative trading optimization strategy for VPPs considering multiple uncertainties is proposed. First, the distributionally robust optimization model for multi-stakeholder collaborative trading is constructed based on the Wasserstein distance. The uncertainties of renewable energy output are represented through distributionally robust chance constraints, and the model is restructured using strong duality theory and the worst-case lower-bound method. Then, a generalized Nash equilibrium model for multi-stakeholder collaborative trading is established. By defining the equilibrium state of the game through a stationary point method and applying linearization techniques, the problem is transformed into a mixed-integer linear programming formulation. Finally, numerical results demonstrate that the proposed collaborative trading optimization strategy effectively ensures reasonable profits for all stakeholders in the VPP while balancing economic efficiency and conservatism.