During the process of developing product, AO directly affects assembly efficiency and assembly quality for mechanical product. Usually, new mechanical product is developed according to the knowledge and experience. So it is very important how to utilize the knowledge and experience to plan the AO. This paper presents a planning approach based on CBR for AO of mechanical product. Firstly, a method based on hierarchy model represents AO cases. And we store the AO cases into the AO Case Library to accumulate the knowledge and experience of assembly process. Secondly, the Euclidean distance is applied to calculate the similar characteristic between the component of AO of Query (AOQ) and the component of the AO cases in the Library. When some distances are less than the value, the corresponding AO cases are selected to the next calculation. Thirdly, the Manhattan distance is applied to calculate the similar characteristic between AOQ and the selected AO cases. When the distance is minimal, the optimal AO case is obtained. The approach is applied successfully to plan AO for a part of aircraft’s wing.