The embodiment of the invention provides a site selection and constant volume optimization method and system for distributed power sources. The method includes the steps that an initial group with the individual number being 2n is generated with power distribution network structural information and capacity information of the positions of the distributed power sources as individuals; n individuals in the initial group are taken as a search group, all the individuals in the search group are subjected to load flow calculation according to improved particle swarm optimization, the fitness of each individual is obtained according to comprehensive evaluation indexes, the position information of each individual is updated so that an offspring group can be generated, and after the predetermined algebra offspring group is generated, the position information of the optimal individual in the last generation is output; the other n individuals in the initial group are taken as a detection group, random iteration is performed on the detection group according to improved particle swarm optimization, and after the number of predetermined iteration times is achieved, the historical optimal position information in the detection group is output; according to the position information of the optimal individual and the historical optimal position information, a final optimization configuration scheme is obtained.