Abstract: After adding prefect photovoltaic power prediction algorithm,the photovoltaic power generation system with hybrid energy storage,which is optimized,can let the photovoltaic power and invert power to keep balance in a short period of time.Then it can reduce the number of storage units frequently depth charging and discharging and prolong the service life of the storage unit and raise the utilization rate of the energy of the system.The existing prediction algorithm of photovoltaic power,which includes BP neural network,similar day typical of the trend and least squares support vector machine etc,has some shortcomings which the utility is not strong and smooth effect is not good.This paper proposes a pratical short-time photovoltaic power prediction mathematical model which is based on the trend of moving average model,the model is based on the historical datas as input,then manipulation datas to get the predicted value.Through the experiment under weather environment of the cloudy, the photovoltaic power prediction algorithm,which can timely tracking and smoothing PV fluctuation,also reduce the storage unit's capacity,is verified.