Abstract: A long and short term planning-running double-layer chance constrained optimization model was proposed for grid connection of distributed photovoltaics（PVs） considering the sequential feature and correlations in active distribution network.The lower short-term model considers operation optimization of the active distribution network in each period to determine the best active management strategy.The upper long-term planning was based on the lower short-term optimization and takes into account the overall economic efficiency of the active distribution network in the whole planning stage to determine the location and capacity of the PVs installation.An improved correlation matrix method with the latin hypercube technique was proposed for describing the correlations as well as the sequential probability distribution differences between PVs and loads,and for dealing with sequential feature and correlation of PVs and loads by time division.Mixed intelligent algorithm combining probabilistic power flow method and nested simulated annealing particle swarm optimization（SAPSO） was employed to solve the established model.The IEEE33 system was selected for example,and the simulation results verify the rationality and validity of the proposed method.