在分析经典模糊时间序列预测模型的基础上,指出了传统的模型不能处理多因素的情形;然后分析并改进了证据理论中关于证据合成的方法,提出了基于证据理论的多因素模糊时间序列预测模型;最后用1997年～2006年10年间的上海股指数据对所提出的模型进行了实践检验,实验结果表明该模型是可行的,其预测效果优于所参照的预测模型. Based on the analysis of classical fuzzy time series forecasting model, the paper points out the traditional model can not deal with multifactorial situations. Then, the paper analyzes and improves the evidence synthesis theory, and proposes a multifactor fuzzy time series forecasting based on evidence theory Finally, the proposed model is tested by using the Shanghai stock index data from 1997 to 2006, and the experimental results show that the model is feasible and the forecasting effect is better than the forecast model.