Based on the antibody clonal selection theory of immunology, an artificial immune system algorithm, clonal selection algorithm based on anti-idiotype (AICSA), is proposed to deal with complex multi-modaloptimization problems by introducing the anti-idiotype. This algorithm evolves and improves the antibodypopulation through clonal proliferation, anti-idiotype mutation, anti-idiotype recombination and clonal selection operation, which can perform global search and local search in many directions rather than one direction around the identical antibody simultaneously. Theoretical analysis proves that AICSA can converge to the global optimum. By introducing the anti-idiotype, AICSA can make the most of the structure information of antibodies, accelerate the convergence, and obtain the global optimization quickly. In experiments, AICSA is tested on four different types of functions and compared with the clonal selection algorithm and other optimization methods. Theoretical analysis and experimental results indicate that AICSA achieves a good performance, and is also an effective and robust technique for optimization.