With the development and popularization of distributed photovoltaics（PV）, virtual power plants have been widely applied as a novel form of electric power system integrating distributed energy resources. However, the uncertainty characteristics of distributed PV output are not fixed but depend on the aggregation scale. Firstly, virtual power plant planning and scheduling are investigated based on the decision dependent uncertainty（DDU）in distributed PV aggregation scale. Then, models for DDU and spatial correlation of distributed PV are established. An optimization risk scheduling model for virtual power plants considering DDU of distributed PV is proposed. Addressing the coupling between decision-making and uncertainty, a solution method based on affine function for the stochastic model of decision dependence probability distribution is presented. Lastly, case studies analyze the impact of DDU of distributed PV on virtual power plant scheduling results and different risk preferences of virtual power plant strategies, verifying the effectiveness of the proposed models.