This thesis investigates the application of evolutionary computing techniques in the synthesis of electronic circuits. Evolutionary computation encompasses a particular class of algorithm which employ some aspects of natural evolution as metaphors. Particularly, most of these algorithms borrow ideas from the natural selection, genetic material recombination and mutation biological mechanisms. Even though evolutionary algorithms have been intensively investigates recently, starting a new research area called Evolutionary Electronics. This work focuses on evolutionary electronics from a enginnering perspective and the main objective is the proposal of a new generation of a Computer Aided Design (CAD) tools. Many case studies have been analysed, covering digital and analog microelectronics. The work aimed the achievement of competitive results comparing to other CAD tools. The research has made use of evolutionary algorithms tailored to these application, by including other genetic operators besides the ones defined above. The following methods have been embedded in the evolutionary methodology: memory based genetic algorithms, use of variable length representation systems and the use of the biological speciation metaphor. Furthermore, a new multiple-objective optimization method, based on artificial neural networks learning algorithms, has also been employed in the case studies. A large number of circuits of practical interest have been sysnthesised, such as filters, amplifiers, oscillators, rectifiers, receptors, comparators refer to new approaches for circuits evaluation, particularly in the digital domain. Circuit simulators and analog the reconfigurable circuits have been used as platforms for the evolutionary process. The results show that the circuits synthesided through evolutionary computation are, in some cases, more efficient than the human designed ones. Besides, the evolved circuits are usually quite different from their human designed counterparts, which can contribute to the creation of new design methodologies. The author identified many promising ways of evolutionary algorithms application in analog and digital design, which may, in the future, overcome conventional design in terms of area, speed and power consumption.