A multi-objective evolutionary algorithm cooperated with decomposition mechanism and opposition-based learning model was proposed for solving complex multi-objective optimization problems. Under the framework of multi-objective evolutionary algorithm based on decomposition, the opposition-based learning model was introduced into the algorithm. The model improved the algorithm’s exploitation. During the evolution process, the opposition-based learning model facilitated the local optimization and the differential evolution strategy enhanced the global research for the new algorithm. The opposition-based learning strategy and differential evolution were in coordination to balance its exploration and exploitation. The benchmark LZ09 series of internationally recognized with complicated Pareto sets were adopted to verify its effectiveness. The proposed multi-objective evolutionary algorithm based on opposition-based learning model was compared with MOEA/D based on DE (MOEA/D—DE), the third evolution step of generalized differential evolution (GDE3), fast and elitist multi-objective genetic algorithm (NSGA—II) and improving strength Pareto evolutionary algorithm (SPEA2), the results showed that the proposed algorithm can obtain Pareto fronts with good convergence, diversity and wild coverage. In order to analyze the algorithm to solve the problem of performance constraints, the proposed algorithm was applied to solve the multi-objective optimization design of speed reducer. The results showed that the Pareto front obtained by the algorithm was uniform, which demonstrated its good performance in solving practical problem with constraints and engineering effectiveness.