Assembly unbalance is a crucial factor in evaluating the assembly performance of aero-engine rotor system, which is influenced by both manufacturing errors and assembly processes. To achieve precise prediction and optimization control of unbalance distribution during assembly process, firstly, this article established a novel APA (Assembly Precision Analysis) model that integrates manufacturing errors and barycenter characteristic based on morphology feature extraction method and homogeneous coordinates transformation theory. Then, taking into account the total unbalance and local maximum unbalance of assembled rotor comprehensively, the dual-objective optimization method for unbalance reduction was proposed. Finally, based on an assembly case of a simulated aero-engine rotor, the application of APA model in optimizing assembly unbalance has been introduced. The analysis results show that the average errors of the APA model in predicting eccentricity value and eccentricity phase are 3.17% and 2.56% respectively. Meanwhile, compared to the existed assembly strategies, the U Total and U Max index obtained by the proposed optimization method are significantly reduced, which proves that the proposed APA model and the dual-objective optimization method are feasible in optimizing assembly unbalance. The research contents have important guiding significance for the high-performance assembly of rotor system.