Abstract: In practical application, especially in the study of large-scale decision space optimization, it is a usual problem that MOEA/D algorithm will fall into local optimization easily. To solve it, a new type of MOEA/D algorithm, which is based on quantum search and Gaussian mutation, is proposed. The quantum search and Gaussian mutation are connected in parallel by means of the environment transfer model introduced. And then the original type of algorithm is connected in series. The quantum search is used to improve the comprehensive search ability of the algorithm, and the Gauss mutation position update method is used to ensure the partial search ability of the algorithm. Meanwhile a quantum search based on neighbor position is proposed to avoid the risk of “precocity” in the later iteration. By changing the generation mode of attraction points, the partial search ability of quantum search in the later iteration is enhanced. The experimental results prove that the improved MOEA/D algorithm, comparing with the original one, has developed the search ability and ensured the convergence ability.