According to the current situation that the level of penetration of intermittent power sources is gradually increasing and the distributed generation(DG) and the capacitor group(CG) are separate planning in the distribution network, the power sources of active power and reactive power can be referred to as generalized power supply(GPS). Output randomness of wind power and photovoltaic power output are considered in this paper. Multi-objective optimal configuration model of generalized power sources is established in distribution system based on chance constrained programming, which takes investment benefit, pollution emissions and branch voltage stability margin as optimal targe. Improved multi-objective particle swarm optimization algorithm-Monte Carlo simulation solves this model. The simulation results of IEEE-33 bus system demonstrate that coordinately Optimizing active and reactive power resources of distribution network can not only improve resource utilization but also reduce pollution emissions and evaluating voltage quality from the viewpoint of probability assists planners in making scientific decisions.