A method of using genetic algorithms to design the optimum passive filter is presented. Attempts have been made for the problems that could not be solved by the classical theory of filter. The result shows that GA is better in adaptability in the design and the problems can be solved in some degree through GA.