To improve the reliability of power supply distribution network, network survivability index is introduced as a new objective function of the distribution network reconfiguration, from the point of network connectivity. At the same time, considering the randomness and intermittent of the output of the wind power generations (WPG), this paper builds up wind power output random variables and their corresponding probability based on wind speed and wind speed probability density function. On this basis, taking the switch state as optimization variables, the output of WPG as random variables, the minimum network loss higher than confidence level as another objective function, a chance constrained programming based distribution network multi-objective reconfiguration model with WPG is built. It applies the improved ant colony algorithm, combines with the strategy of spanning tree which ensure that the ants path to meet the structure of the distribution network, to solve the built model. And then, the Pareto optimality is used to evaluate the result to get the optimal solution set. The results of IEEE33 system and PG&E69 system verify the validity of the model and algorithm.