With the increasing penetration of distributed renewable energy connected to the distribution network, the safe operation and control of the distribution network becomes more complex, which brings new challenges to the reliability analysis of power supply. This paper presents a new power supply reliability analysis method for active distribution networks considering the renewable generation output uncertainties and islanding operation. Firstly, a two state reliability model are established key components of distribution networks and renewable generation equipment, and based on the probability distribution analysis of renewable energy resources, such as wind and solar, a multi-state reliability model for potential renewable power output (REPO) has been developed, which leads to the proposal of a multi-state reliability model for renewable power stations considering equipment conditions. Secondly, the concept and benefit of islanding operation by the feeder partition with embedded distributed generation were analyzed, its impact on power supply and reliability analysis were modelled. Finally, an active distribution reliability analysis method based using sequential Monte Carlo simulation technique is used to assess the active power distribution network power supply reliability. The effectiveness of the proposed method is verified and applied using an IEEE RBTS test system.