The invention discloses a wind power output typical scene generation method suitable for mid-and-long term power planning. The method comprises the steps that historical data are acquired, including the s-season typical daily load curve and the s-season daily wind power output curve; the daily peak period, the daily intermediate load period and the daily valley period are obtained according to the peak, valley and intermediate load characteristics of the s-season typical daily load curve, and then the characteristic index of the s-season daily wind power output curve can be obtained; and the power balance scene and the probability thereof, the peak load balance scene and the probability thereof and the power balance scene and the probability thereof of the random wind power output typical scene can be obtained according to the characteristic index of the s-season wind power output curve. The characteristics of mid-and-long term power planning and the power structure characteristics are comprehensively considered, and the uncertainty, the seasonal characteristics and the source-load interactive characteristics of wind power output are considered so that the method can be used for various occasions of mid-and-long term power planning and analysis of the related topics.