The power outage accidents of transmission line happened more and more frequently because of forest fire disasters, which seriously affected the safety and stability of the grid. Forest fire risk the transmission line is becoming an important research topic for disaster prevention and mitigation of grid. Considering the complexity and variability of fire risk factors, this paper proposes a forest fire risk assessment model based on BP neural network of transmission line. Combining with the actual situation of high incidence of forest fire disasters in some places of 220 kV transmission lines and above, several factors and the actual fire assessment grade are chosen to serve as the input of model and the output of model respectively. Thus, the risk assessment model is established. The experiment results show that the model predicts fire risk of transmission line effectively, which is of great significance to give a warning message timely for grid.