Abstract: Based on the study of the existing evolutionary algorithm (EA) and the immune principle of creatures, an immune evolutionary algorithm (IEA) is proposed. It is a global optimization method, in which offsprings are produced by the best individual of parents population and the elitist replaces population to converge. In the process of optimization, the new algorithm increases its convergence speed, combining certainty with randomicity effectively. A global convergence is proved by analysis of Markov chain. The test results show that IEA not only requires simple parameters but also can improve the convergence.