Based on the analysis of some conventional fuzzy time series forecast models, this paper argues that these
conventional models can not deal with multiple-attribute fuzzy time series and proposes a model based on Dempster-Shafer
theory to overcome this shortcoming with some revise for combining belief function. The Shanghai stock index over 10 years
period 1997∼2006 is chosen as the forecasting targets, and the empirical results show that the proposed model outperforms
the compared models.