Large power transformer is the key equipment of power transmission systems, and its reliability is directly related to the safety and stability of power grids operation. In order to improve the efficiency of the transformer reliability analysis, a reliability tracing analysis method for multi-state power transformers using Bayesian Network is proposed. The unreliability tracking algorithm for multi-state transformer is proposed. According to the relevant references, expert experience and data collection, the components based power transformer Bayesian network model is built. Combined with the reliability index statistics, the key components impacting reliability and the weak parts of transformer are recognized by the proposed technique. Through the use of Bayesian network to track the transformer reliability analysis, the data to support the condition based maintenance and all-life management for the transformer is provided.