Considering the uncertainty and stochasticity of intermittent distributed generations (DGs), a scenario method using Wasserstein distance metric and K-means cluster scenes reduction technique to generate optimal scene is proposed in this paper. So the stochastic optimization problem is transformed into a deterministic optimization problem. The multi-scenario tree models of wind-photovoltaic-load are established. A multi-objective reactive power optimization mathematical model of active distribution network containing intermittent DGS is built, which includes objectives that are the total active power losses and the voltage deviations of the bus, and considering energy storage states of charge characteristic constraints. Also, the artificial bee colony algorithm is used to solve the optimization problems. Simulation results show that the optimal scenes based on the Wasserstein distance indicators and K-means cluster technology reflect the random feature of distributed generation active power output more accurately. Finally, the simulation analysis of IEEE-33 bus distribution test system is carried out to verify the effectiveness and feasibility of the proposed method. This work is supported by National Natural Science Foundation of China (No. 51467009).