Based on the clonal selection theory of the immune system, a novel optimal algorithm, called a directional immune clonal selection algorithm (DICSA), is proposed to design RF diplexer. In DICSA, antibody is proliferated and divided into a set of subpopulation groups. In the antibody's updating, the clonal mutation is applied to avoid premature convergences. The clonal selection realizes the information communication between the subpopulation groups so as to improve the search efficiency. DICSA is applied to a practical problem, designing RF diplexer. The results show that DICSA has stable performance during the optimization process with a satisfactory result.