The power mix composition is really important for the price, security and environmental aspects. The different points of view have different optimization objectives. In this approach we investigate the security extended by risk problems. First we assembly a set of risk of the power generation then we build up a risk matrix of the possible set of the generators. After weighting the importance of the risks we write the total risk of the power mix as linear composition of risk and possible feeding in power. By a linear constraint programming we look for the optimal power plant mix. Finally this mix is compared with the results of former Single Objective Optimizations. A measurement number is introduced to compare distinct power mixes. The calculations show that the risk based power mix composition results a quite modest and balanced usage of the available generators.