In recent years, sensor-kind network devices have been deployed in many applications,

which can generate a lot of raw data continually. However, the raw data is of high level of uncertainty because of the inherent unreliability of sensor readings. Meanwhile, the rules used to model the real world are simplified so that they do not capture the complexity of application scenarios and introduce the uncertainty. Consequently it is of great value to model uncertainty in complex event processing. In this paper, we survey the uncertainty in CEP to help researchers in understanding different approaches and mechanisms discussed so far, according to the classification of uncertainty for CEP from uncertainty in events to in rules. In particular, we propose a general framework to deal with uncertainty in complex event processing and look into the future about the research work.