Based on the complex network theory, the structural vulnerability of a distribution network was assessed from the view of topology of structures. A multi-objective planning model for distribution network optimization was built, with aims to minimize the summation of investment and operation cost, as well as to optimize the structural vulnerability index. An improved minimum-cost spanning tree algorithm was adopted to optimize this planning model. By using the sum of the decrement rate of the economic target and the increment rate of the structural vulnerability target as the objective functions in the iteration process, the multi-objective problem could be solved as a single-objective problem. Results of a sample system with 16 nodes verify the feasibility of the proposed method used for multi-objective structure optimization of the distribution network. The developed planning model can effectively evaluate the structural vulnerability of a distribution network, so as to acquire an optimal planning scheme with a stable structure and with little cost.