The invention provides an ultrashort-term prediction method and system for wind electricity power. The method comprises the steps that a local extremum point is detected in a main component fluctuation sequence of the wind electricity power, and an extremum point sequence of a continuous extremum variation range under the constraint of a preset threshold value of an installed capacity; selecting an adjacent extreme point of sampling data from the extremum point sequence, and defining a historical complex fluctuation sequence; calculating an Euclidean distance between the historical complex fluctuation sequence and a predicted complex fluctuation sequence at the next moment; conducting descending sorting on the historical complex fluctuation sequence according to the Euclidean distance,anddetermining a serial fluctuation sequence according to fluctuation characteristics; conducting serial fluctuation tendency fusion on the serial fluctuation sequence, and obtaining an ultrashort-term prediction result. Through the utilization of the statistical characteristics of the steady-state tendency of the main component fluctuation of wind electricity, in combination with future fluctuationsituation judgement, the high-precision ultrashort-term prediction of the wind electricity power is achieved.