The invention discloses an electrode critical coronal voltage predicting method on the basis of characteristic quantities of electric fields, which comprises the following steps: firstly, measuring critical coronal voltage data of typical electrodes which are of different structures and defining a non critical coronal voltage interval and a critical coronal voltage interval; then carrying out electric field calculation on gaps between the typical electrodes which are of different structures and extracting the characteristic quantities of the electric fields so as to construct a training sample set; nextly, on the basis of the training sample set, constructing a critical coronal voltage predicting model, wherein the critical coronal voltage predicting model uses the characteristic quantities of the electric fields as inputs and uses the non critical coronal voltage interval and the critical coronal voltage interval as outputs; and finally, predicting critical coronal voltages of the electrodes by adopting the critical coronal voltage predicting model. The electrode critical coronal voltage predicting method is simple to operate, has high predicting accuracy and is suitable for predicting the critical coronal voltage of the electrode which is of a random structure; and a predicting result can provide reference for structural design and structural optimization of a power transmission line, fittings, an electrostatic dust collection device and the like.