随着Web Service技术的迅速发展,具有相同或相似功能的服务在互联网上大量出现.如何预测服务质量,进而支持基于服务质量的服务选择,成为了需要解决的关键问题.其中,响应时间是服务消费者最能直接感受到的服务质量体现,提出一种Web服务响应时间的动态预测方法.该方法根据响应时间的历史数据,基于时间序列分析进行动态预测.实验结果表明,这种方法可以显著提高预测的准确性,更好地支持服务的选择. With the rapid development of Web Service technology, many services with the same or similar functions appear on the Internet.How to predict the quality of service, and then to support the selection of service based on quality of service, has become a key problem to be solved, in which the response time is service This paper presents a dynamic forecasting method of Web service response time which is most directly perceived by consumers and proposes a dynamic forecasting method based on the historical data of response time based on time series analysis.The experimental results show that this method can be significantly Improve forecast accuracy and better support service choices.