Abstract—Currently, the safety monitoring process in a complex plant environment is proceeded by the human. Sometimes, human error that may occur in a filed causes a severe problem. This paper introduces a soft-computing method for detection of the abnormal status of plant equipment using sound information and neuro-fuzzy theory that is one of the intelligent theories. The sound for testing is acquired from the cylinder valve, compressor operation, safety valve open. In this paper, the high-pressure gas filling plant will be used as a test plant. The resulting system will be widely applied to more complex plant environments.