The integration of distributed power sources and electric vehicles introduces significant stochasticity to the distribution grid. In order to better adapt to real-world applications, a model for uncertain scenarios considering the uncertainty of sources and loads is established, with the total operating cost, voltage deviation, and voltage stability as objective functions for distributed power optimization. Firstly, probabilistic models are developed for distributed power sources and electric vehicles. Secondly, Latin hypercube sampling is utilized to generate initial scenarios, followed by the utilization of the k-means clustering algorithm to reduce the initial scenarios. Finally, an improved grey wolf optimization algorithm is employed for optimization. Based on the IEEE33-node standard test system, simulations are conducted. Besides, it can be founded that the obtained fitness value of the proposed method is minimized compared with particle swarm optimization and grey wolf optimization algorithms. At the same time, the rationality and effectiveness of the proposed model and algorithm is further validated.