The invention discloses an uncertainty analysis method and system for wind power plant prediction. The method comprises: obtaining turbulence intensity levels from obtained wind power plant data according to a preset division threshold value; constructing a conditional quantile regression model based on the turbulence intensity level, the actually measured power and the predicted power; and combining the turbulence intensity levels and a conditional quantile regression model under each turbulence intensity level to obtain the uncertainty of the predicted power. On the basis that uncertainty isdetermined through a quantile regression method, a conditional quantile regression model is established according to research on statistical characteristics of turbulence intensity variables under all turbulence intensity levels, and therefore an original wind power prediction uncertainty analysis method is partially optimized.