The aggregated power of distributed PV at the point of common coupling (PCC) on the AC side has large randomness and volatility, affecting the stable operation of the grid. To help combat this, an adaptive PCC power fluctuation smoothing method based on empirical wavelet transform (EWT) is proposed. First, this paper focuses on a typical scenario of a hybrid energy storage system (HESS) and distributed PV access to the PCC. After analyzing the characteristics of the EWT adaptive processing waveform, and combining with the power fluctuation rate and energy storage element response characteristics, the EWT decomposition and optimal correction of the original PV aggregated power are applied to achieve HESS primary power distribution. Then to avoid the HESS state of charge (SOC) from frequently crossing its limit, an active power compensation control is proposed to make it work within the safe range by actively changing the energy storage reference signal. A simulation example combined with actual data shows that the smoothing method can adaptively realize reasonable decomposition and power distribution of PV output and effectively meet the requirements of grid-connected power fluctuation while extending the service life of energy storage. This provides a new idea for smoothing PV output power fluctuation.