In order to solve the shortcomings of existing distribution system reliability dependent simulation methods and imprecise optimization methods, a new and efficient distribution system reliability evaluation method based on algebraic model is proposed in this paper. Firstly, four types of commonly-used distribution system reliability indicators are constructed by using the expected node outage rate and duration, and then, they are reconstructed into an equivalent recursive algebraic model based on node related information. Finally, the topological related information is transformed into algebraic constraints and embedded into the recursive algebraic model, which is reconstructed into a group of algebraic models, so that the reliability index can be calculated efficiently and conveniently. The simulation results show that the calculation speed required by the algebraic method proposed in this paper is 2~3 orders of magnitude faster than that of the reliability evaluation method based on Monte Carlo simulation and the reliability evaluation method based on optimization, respectively, which is conducive to further application in distribution network planning and operation.