The invention discloses a method for predicting a 5G millimeter wave network signal strength spatial distribution situation. The method comprises the steps of carrying out inversion to obtain dielectric parameters on surfaces of building materials in a prediction area; establishing a path plane model and a sight distance plane model, giving a maximum multipath distribution result for a special transceiving end location and selecting valid reflection multipaths; and carrying out direct wave consistent area division on the prediction area, measuring direct wave field intensity amplitudes and phases of direct wave consistent areas, predicting field intensity of certain receiving point, judging the direct wave consistent area to which the certain receiving point belongs, calculating a direct reflection field of the receiving point according to a field intensity measuring value of the direct wave consistent area, judging the direct wave consistent area to which a reflection point belongs according to reflection point locations of the valid reflection multipaths, calculating a reflection field by taking a transmission field amplitude of the direct wave consistent area to which the reflection point belongs as an incident field amplitude of a reflection wave, overlapping the direct field and all reflection multipath fields, determining an electric field vector of the receiving point, and summarizing the receiving field of each point in the prediction area. Electromagnetism situation distribution is predicted.