Aiming at the problems of insufficient precision in fault identification and the lack of decision-making rationality in transmission line reclosing during winter, a hierarchical reclosing strategy based on fault cause identification and multi-source prior knowledge is proposed. First, a fault transient waveform feature extraction method based on time-frequency network (TFN) is constructed, and the causes and probabilities of typical winter faults such as icing, wind swing, wildfires, and foreign object interference are identified by integrating weather information. Second, the beta-binomial distribution is used to estimate the parameters of historical reclosing prior information, capturing its inherent uncertainty. Then, the fault cause identification results are combined with sampling values from the reclosing success probability distribution, and a Bayesian conditional probability model is used to calculate the reclosing success probability under current fault conditions. Based on this, thresholds are set for different operating scenarios, forming a hierarchical reclosing decision-making mechanism. Finally, case studies on actual transmission lines show that the fault cause identification accuracy of the proposed method reaches 95.80%. Compared with the existing reclosing success rate of 52.96%, the proposed hierarchical reclosing strategy achieves a decision accuracy of 79.41% and reclosing success accuracy of 82.52%, significantly improving the reclosing success rate and reducing the impact of blind reclosing on the power grid.