To carry out automatic path decision-making for the penetration of Unmanned Aerial Vehicles (UAVs) in threat-networked environment, a model for path decision-making is proposed based on Structure-Varied Discrete Dynamic Bayesian Network (SVDDBN). The decision-making model is established according to the mode of UAV passing through the threat area and its response to firepower threat. Then the results of path decision-making are obtained by using the SVDDBN algorithm based on adaptive parameter. The results of simulation show that the model meets the need of path decision-making for UAVs in threat-networked environment and provides a new pattern of thinking for the penetration of UAVs.