A surveillance system and method are provided. The surveillance system includes an audio processing device which extracts an audio feature of an audio signal, and determines whether an abnormal event has occurred in a monitoring region, based on the audio feature; a video processing device which extracts a foreground region from a video signal, and determines whether an abnormal event has occurred in the monitoring region, based on motion information of the foreground region; and a context awareness device which calculates an audio abnormal probability and a video abnormal probability by respectively accumulating results of abnormal event occurrence/non-occurrence determinations performed on audio signals and results of abnormal event occurrence/non-occurrence determinations performed on video signals for a certain period of time, and finally determines whether an abnormal situation has occurred in the monitoring region, by using respective combined probability distribution models for a normal situation and the abnormal situation.