In order to firstly assure the mating features of high importance during the selective assembly process, a selective assembly method considering the feature weightings is proposed. All of the features are evaluated by fuzzy entropy weight method. Their weight coefficients which present the importance are defined. Based on the dissymmetrical quality loss function and the weight coefficients, the multi-objective model for selective assembly is established. A specific generic algorithm considering target weightings is used to solve the optimization model. It can search the preference region using the Pareto dominant order to derive satisfactory solutions with high quality. Finally, the method is used for the selective assembly of a certain aero engine’s subassembly to invalidate the feasibility.