With the development of radio frequency identification (RFID) technology, RFID data processing technique grows rapidly in demand in the multi-context aware environment. According to the characteristics of RFID data, the event-centric event model is built. Composite events are abstracted by analyzing and processing the raw RFID data. And then the abstract low events can be monitored by RFID event detection. With the base of event specifying and detecting of RFID data in Cascadia, the integrated event operators and the main idea of detection algorithms are given based on event parsed tree. The detailed event detection algorithms are presented in the different constraint executing circumstances. The theoretic analyses of the algorithms are also given accordingly. Furthermore, the RFID composite event detection model using event parse graph is explained by the applied examples. The experimental results indicate the efficiency and effectiveness of the algorithms in three varied constraint environments in RFID data management.