Molecular devices are potential candidates for the next step towards nanoelectronic technology, and they would make it possible to realize the most advantageous applications. Since Aviram and Ratner suggested that a unimolecule with a donor-spacer-acceptor structure would behave like a diode, a great deal of experimental work has been carried out to demonstrate an element of such an electronic device using either a single- or a finite-number of small organic molecules. The development of molecular nanoelectronics necessitates precise control based on the atomic and/or molecular scale. An important process for realizing this technology is the design of molecules for specific functions, such as wires, diodes and transistors. In this research area, the design of molecules with desirable properties for molecular devices is still a great challenge, since the behavior of the molecules is largely unpredictable. Therefore, computational methods can be used to assist and accelerate this survey process. However, the number of possible combination is very nearly infinite, and we cannot survey all molecules using a computer simulation, even if those molecules have already been synthesized. In this paper we propose a Genetic Algorithm (GA) method, which can be used to survey appropriate materials for a molecular wire, and we describe molecules that can be obtained from simple oligomers, for example, acetylene, phenylene, thiophene oligomer, and so on. This technique seeks to mimic natural evolutionary principles (Darwin's principles) to identify promising functional molecules. We discuss an evolutionary process to escape a local minimum in the quest to find the most appropriate molecule.