In recent years, extreme disasters have been the main threat to the power distribution network. Allocation of disaster prevention resources before the disasters is of great significance to promote the power distribution network’s resilience and guarantee continuous power supply of load. However, current allocation strategies have not fully considered the outage risk of each node. This paper utilizes the Bayesian network to calculate the outage probability of each node, and according to it, this paper proposes a pre-disaster resource allocation strategy considering load capacity, importance level, and outage risk. Furthermore, an optimization model for allocating resources has been formulated. By case analysis, the strategy proposed by this paper fully consideres the indices, including outage risks, and reduces the outage risk of the whole network, and thus the resilience of the power distribution system has been promoted. The effectiveness and robustness of the strategy have been validated.