To improve the forecasting precision of electricity consumption, and content the demand of marketing, we establish the model by combining particle swarm optimization (PSO) with the improved grey theory. The method is simple, easy to do practice and its convergence rate is quick. It can find the overall optimal solution of problems in great probability, and can effectively overcome the shortage of the traditional grey theory. To prove the validity, this paper forecasts the electricity consumption in Bayannur region. The result shows that, the model has higher forecasting precision and extensive application.