The invention provides a photovoltaic power station power prediction method and device, and the method comprises the steps: obtaining meteorological data predicted for an area to which a photovoltaicpower station belongs, and extracting meteorological features from the predicted meteorological data; inputting the meteorological characteristics into a neural network model constructed for the photovoltaic power station to obtain generated power output by the neural network model, and taking the generated power as a generated power prediction value of the photovoltaic power station, wherein theneural network model is obtained by training historical meteorological data of an area to which the photovoltaic power station belongs and historical generated power of the photovoltaic power stationcorresponding to the historical meteorological data. According to the invention, the prediction of the generated power of the photovoltaic power station can be realized through the above mode.