Abstract: To improve the optimization quality of cloud manufacturing service composition, an optimization method for multi-objective cloud manufacturing service composition based on improved NSGA-II algorithm was proposed. Firstly, improved the concept of dominant strength was quickly determined the ordering of individuals in non-dominated set. Then, different local search strategies were applied to the NSGA-II algorithm to accelerate the convergence by strengthening the search for excellent individuals in the early stage of the algorithm. In the later stage, sparse individuals were searched by using method combining neighborhood search and simulated annealing algorithm to increase population diversity. Finally,the effectiveness of the optimization model and the feasibility of the algorithm are verified by an actual enterprise case．