Slow convergence to the global optimum has been one of the main problems in genetic algorithm. In order to increase the speed of convergence, an immune genetic algorithm based on vaccine autonomous obtaining and updating (IGAVAOU) is propos ed. Excellent individuals are selected from each generation population and vacci ne is obtained from these excellent individuals. Then individuals in succeeding population are vaccinated in stochastic way. Vaccination is a kind of operation by which allele in vaccine replace allele on individual corresponding locus. Vac cination can not only make excellent schemata proliferate, but also repair the s chemata destroyed by crossover and mutation operations. Population and vaccine r epertory influence each other and co-evolve so that they accelerate convergence to the global optimum. IGAVAOU's computation efficiency is analyzed based on the schemata theorem. IGAVAOU is verified by several typical functions. The results show the feasibility and validity of the algorithm.