Service-aware computing is a hot research topic under the banner of web-based uncertain mobile applications. As we know, in the research domain of uncertain mobile service, service-aware evidence with uncertainty is dynamic and changing randomly. In order to ensure the QoS of different mobile application fields based on decision making, we think the method of service-aware computing for uncertain mobile applications is very important. The key insight of this paper is that we modified the computing method of evidence information, which has been considered the reliability, time-efficiency, and relativity of service context. The method has improved the classical computing rule of D–S (Dempster–Shafer) Evidence Theory when being used in uncertain cases. The novel method may be called the extended D–S (EDS) method, which has overcome the drawbacks of classical D–S Evidence Theory. All these new ideas have been successfully used in our service-aware computing field of uncertain mobile applications. By comparing EDS with related methods, such as Bayesian Theory (BT), and Random Set Theory (RST), the advantage of the new service-aware computing method has been proved successfully.