Abstract: At present,the alarm method of home-made computer monitoring system of hydropower station has such problems such as low alarm efficiency,high false alarm rate,and low alarm efficiency,which is difficult to ensure the normal operation of the system. In order to ensure the safe and stable operation of home-made computer monitoring system of hydropower station,this paper proposes a new abnormal behavior alarm method. Based on the time series,the time series data of the collection device is described by constructing a time series symbol conversion model. The morphological characteristics of the device are combined with feature recognition technology and abnormal behavior detection technology to accurately judge the operating state of the device,and the alarm line is preset to realize automatic alarm. Experimental results show that the alarm method has strong adaptability,with an average effective alarm rate of 94%,an average false alarm rate of8%,and an alarm efficiency of about 7.8 units/s. It has high alarm performance and application value.