The need to decrease fuel consumption and environmental pollution caused by ships leads to the development of complex and integrated systems that produce all forms of energy needed (mechanical, electrical and thermal) with as low fuel consumption as technically possible and economically feasible. A variety of configurations, design specifications and operating states can serve the loads. The question of which combination of those is the best under certain criteria is far from trivial, and relevant engineering decisions have to be supported by optimization procedures. As an aid for this purpose, a method has been developed and presented in this article. A superconfiguration of the system is considered and simulation models of the system and its components are developed. Instead of the three-or two-level approach followed by other methods for the solution of the triple optimization problem (synthesis-design-operation), a single-level approach is followed by the present method. As for the operation optimization, all the operating modes are optimized simultaneously. Due to the discontinuity and multimodality of the optimization problem, gradient-based methods are not particularly successful for the solution. Instead, a genetic algorithm is used. The applicability of the method is demonstrated with numerical examples, while its versatility, i.e. its adaptability to alternative conditions or values of parameters, is demonstrated by solving the problem with different sets of data, such as varying fuel price, capital cost and operation profile. Even though the application example is for a ship energy system, the method can also be applied for the optimization of land installations, such as power plants and cogeneration systems, with proper modification.