Abstract: In this paper, research status of distributed generation (DG) planning is reviewed and a two steps approach of DG sitting and sizing is proposed to break the limitation of DG capacity. The solution is to site and size new energy DG, based on which other kinds of DGs can be optimally allocated. By considering of reducing the network operation cost and improving voltage quality, the objective is to reduce equivalent carbon dioxide output of power system, so the difference of cost and income before and after DG planning is taken as the objective function, which reflects the economic benefit of distributed network intuitively. Taking the IEEE 33-bus system as an example, the self adaption particle swarm optimization (SAPSO) algorithm is used to optimize allocation scheme of DG. In addition, the node and capacity of all DGs are given by graphs and tables, which show that the portion of photovoltaic generation is 60% of total distributed generation. The indexes caculation shows that power loss, voltage quality and the emission of carbon dioxide are improved greatly.