The invention discloses a photovoltaic power prediction method and system based on an approximate period time series. The method comprises the following steps: acquiring historical data of photovoltaic power generation on a day, and preprocessing the historical data to obtain a daily power curve; identifying the approximate period of the daily power curve and generating an approximate period sequence; according to the approximate periodic series, obtaining the time transform function of the cycle period, and obtaining the prediction time to reach the peak value in the next cycle; outputting apower generation prediction result of the next cycle. The method and system are developed based on the approximate periodic time series, and the seasonality and trend law of the method and system arein high agreement with the photovoltaic power generation curve, and the simulation effect is better. At the same time, the influence of weather factors on forecasting is considered comprehensively, sothat the forecasting results are more reasonable and accurate, and the forecasting method is effective and feasible, and accords with the reality.