Abstract: It is an inevitable trend to develop and refine the evaluation index of smart grid. Being one of the main stakeholders, to improve users ability of smart electricity consumption is an important part in smart grid. Therefore, the household smart electricity consumption index system is built based on such principles as systematicness, typicality and comparability. At first, the method of expert group decision is used to supplement the index set, which enhances the comprehensiveness of the original system. Then mathematical programming method is applied to estimate the deviation between the original judgment matrix and the ideal value, and a kind of interval calculation model for index weights that reflects experts decision bias is proposed by combining with experts credibility. In addition, the weak weight index is eliminated to improve the conciseness of the index system. In the end, the index is applied to evaluate the actual community, whose results show that the optimized index system can effectively and reliably evaluate the level of household smart electricity consumption.