Abstract: Analyzing the model of the single level capacitated lot-sizing problem (SLCLSP), a new approach based on IGA (immune genetic algorithm) is proposed to solve the problem. It not only retains the global research ability of SGA (simple genetic algorithm) but also adopts some strategies to keep up the diversity of biological immunity mechanism so as to improve greatly the diversity of population of SGA. Experimental results showed that the proposed IGA can rise above efficiently such difficulties of SGA as precocious convergence and poor local search ability and provide well the global converging ability'to enhance both the global convergency and convergence rate, thus solving effectively the SLCLSP problem.