Acquisition granularity and acquisition span are two important indexes to analyze the active power of renewable energy power stations, and it is important for the analysis of the intermittent energy output power to determine the acquisition granularity of the data. An acquisition granularity calibration method of the photovoltaic output power based on data mining technology is proposed in this paper. The sensitivity changing characteristics of the time series and the sum of the output power are respectively extracted by the multi-scale descriptive statistical analysis and interpolation method. A “power trapezoidal continuous changing state” method is proposed to establish a multi-objective optimization model for the acquisition granularity calibration of the photovoltaic output power. Genetic algorithm (GA) and Particle swarm optimization (PSO) algorithm are respectively used to solve the model and determine the optimal acquisition granularity of the photovoltaic output power. The sensitivity of the acquisition granularity of the data to the capacity of the energy storage system is analyzed, and the energy storage system with the optimal acquisition granularity can’t only effectively smooth the fluctuation of the photovoltaic output power but also keep the main information of the data. The simulation tests of the annual actual operation data at a photovoltaic power station with the installed capacity of 14MW in China verify the validity of the model. The simulation results show when the acquisition granularity of the photovoltaic output power takes 45s, it can satisfy the accuracy of the required data for the capacity configuration of the energy storage system. The method proposed in this paper provides a theoretical basis for the intermittent energy applications and has a certain engineering application prospects.