V output is intermittent and volatile, and the use of energy storage devices can reduce the impact of PV grid-connected on the grid. A capacity configuration of a hybrid energy storage system (HESS) composed of super capacitor and battery is proposed based on the effect of smoothing the PV power fluctuation. An algorithm of recursive average and median average weighted filtering is used to smooth the PV output, and the PV grid-connected power and HESS reference power are obtained. The high frequency and low frequency components of HESS reference power are extracted by complete ensemble empirical mode decomposition (CEEMDAN) and Hilbert transform (HT), and the internal power of HESS is allocated once. Then, in order to avoid frequent over-limits of HESS state of charge (SOC) and to ensure that HESS can continue to work, this paper proposes a balanced management method of HESS power. This corrects the attenuation of primary distribution power, and then configures the capacity of the corrected HESS reference power. Finally, an economic evaluation model is established based on life cycle cost (LCC), and different energy storage capacity allocation strategies are adopted. The effectiveness and economy of the proposed method are verified by actual data simulation. This work is supported by the National Key Research and Development Program of China (No. 2016YFB0900600).