Abstract: Aiming at the deficiency and inaccuracy in multi-objective optimization for assembly sequence planning, a novel assembly model and optimization algorithm is proposed. A more compact contact-connection matrix is constructed to provide more information in relation to assembly, and an extended interference matrix is given to solve the assembly problems in non-orthogonal direction to which other existing methods are unable. The methods for automatically acquiring the matricies are also studied. Several rules that affect assembly efficiency are studied from the viewpoints of automatic planning and quantification, with the flexibility provided for customizing them and their priorities. Then, an optimization algorithm is proposed via screening out the information as above stepwise so as to form a wholly optimized sequence. A prototype system is developed on UG platform with a valve taken as example to verify the feasibility and effectiveness of the method proposed, as well as its good visualized simulation effect.