Energy storage can smooth out the output power fluctuation of distributed generations and potentially improve the supply reliability of distribution system. To quantitatively assess the effect of energy storage， this paper proposes a pseudo-sequential Monte Carlo simulation method for reliability evaluation of distribution system including distributed wind turbines， photovoltaic arrays and batteries. The hourly sequential models and state variation models of wind turbines， photovoltaic arrays and batteries are established. On this basis， the non-power components are sampled sequentially and the components of wind turbines， photovoltaic arrays and batteries are sampled non-sequentially. The failure modes and effects analysis procedure of distribution system is discussed and the reliability evaluation flow is illustrated. Taking the modified IEEE RBTS as an example， system reliability indices in different situations are compared and analyzed， and the effectiveness of the proposed method is verified.