A concept of optimizing the grid connected new energy is proposed to cope with the bilateral randomness problem of power system. In illusion to the problems such as excessive investment, voltage drop and the ascending loss of active power which are resulted from the differences in cost, revenue and capacity between various kinds of new energy. A mathematical model for multi-objective optimizing with the maximum rate of return on investments, the minimum power loss and the offset of node voltage as its objective is established which is based on the Genetic Algorithm (GA) with the elitist strategy and multi-parameters. This paper reaches Pareto optimal solution space of new energy decisions and the overall optimal strategy through normalized processing. Simulating analysis conducted with the instance of IEEE-30 node system verifies the rationality and validity of this method which could provide certain reference for the equipment selection for the new energy on the supply side.