The optimization of energy structures, aimed at saving energy and reducing emissions, is an important precautionary measure against climate change. This study considers different environmental impacts of power systems, and investigates ways to optimize power structures and decrease their potential environmental impact. A multi-objectives optimization model of energy structures was created based on life cycle assessment (LCA). This model covers several environment impacts, rather than only focusing on carbon emissions. LCA was used to calculate the different environmental impacts and provided a new method for normalization. The model was applied to the power industry in China. Three kinds of environmental impacts were considered: material input (MI), global warming potential (GWP), and water deprivation (WD). The five major existing methods of electricity generation in China were considered: thermal power, nuclear power, hydro power, wind power, and solar photovoltaic power. The system boundary included all life cycle stages; specifically, extraction of raw materials and resources, production, energy generation processes, and power transport. The optimization results showed that the total environmental impacts were reduced; MI, GWP, and WD were decreased by 29.53%, 29.67%, and 19.06%, respectively. This method provides new insights into optimization of energy structures by considering multi-environment impacts.