The invention relates to an exponential smoothing method-based power equipment alarm information trend prediction method and belongs to the power system dispatching automation technical field. According to the method, the quantity of the historical alarm information of each transformer substation is calculated, the quantity of the alarm information of each transformer substation per day is calculated and is adopted as an observation value; the initial value of an exponential smoothing method is determined according to the number of days of historical alarm information observation; a smoothing coefficient is selected according to a historical data trend; a prediction model of single exponential smoothing is built so as to be used for predicting the possible value of the quantity of the alarm information of each transformer in the future, and therefore, the burden of monitoring personnel can be decreased; the historical alarm data can be well utilized; the historical alarm data are mined and analyzed; the content of monitoring applications is enriched, and the practicability of the monitoring applications is improved; equipment operating trend early warning capability can be improved; and the leap of security protection from post-event analysis to pre-event predictive control can be realized.