In view of the insufficiency of decision-making ability in the existing distributed generator（DG）optimal planning method，a model for DG optimal planning is proposed along with its algorithm with information uncertainty taken into account.Under the condition of uncertain type，location and capacity of DGs，a multi-objective and non-linear optimal model is developed with the least investment cost，the highest revenue，the highest environment benefits and the minimum loss in distribution network as the sub-objectives.For cases in which there exist many uncertainty factors，plus their hard-to-define probability distribution，interval numbers are used to deal with the uncertainty of the property values and their weights in the actual decision-making of DG optimal allocation.A method by combining the technique for order preference by similarity to ideal solution（TOPSIS）algorithm based on interval mathematics and genetic algorithm（GA）is proposed to solve the problem.The feasibility of the method proposed is shown by a typical distribution network test example.