As an aggregator of distributed energy resources (DERs) such as distributed generator, energy storage, and load, the virtual power plant (VPP) enables these small DERs participating in system operation. One of the critical issues is how to aggregate DERs to form VPPs appropriately. To improve the controllability and reduce the operation cost of VPP, the complementary DERs with close electrical distances should be aggregated in the same VPP. In this paper, it is formulated as an optimal network partition model for minimizing the voltage deviation inside VPPs and the fluctuation of injection power at the point of common coupling (PCC). A new convex formulation of network reconfiguration strategy is incorporated in this approach which can guarantee the components of the same VPP connected and further improve the performance of VPPs. The proposed approach is cast as an instance of mixed-integer linear programming (MILP) and can be effectively solved. Moreover, a scenario reduction method is developed to reduce the computation burden based on the k-shape algorithm. Numerical tests on the 13-bus and 70-bus distribution networks justify the effectiveness of the proposed approach.