The embodiment of the invention provides a line hidden danger prediction method and device, and is used for predicting tree growth under an overhead line. The method comprises the following steps that: obtaining the historical growth data, the growth environment data and the weather data of trees under the overhead line; taking the growth environment data and the weather data as independent variables and the historical growth data as a dependent variable to train a growth prediction model, and obtaining the growth rhythm of the trees under the overhead line; and according to the growth rhythm,obtaining a time node when the tree growth under the overhead line is a hidden danger. Compared with the prior art, the method improves hidden danger prediction accuracy, improves the precognitive ability of operation and maintenance personnel for tree hidden dangers and lowers the occurrence probability of the class of faults while a human inspection frequency is reduced.