The invention relates to a human-vehicle identification method based on four-dimensional information weight, and belongs to the technical field of radar target identification. The method comprises thefollowing steps: 1, caching tracking data of a target for X times; 2, calculating X normalized RCS values; 3, analyzing the spectrum of a radar target echo after MTD, finding out the maximum spectrumpeak position of the target, and calculating the number of other spectrum peaks of the target echo in the +/-N point of the position; 4, calculating a normalized RCS first-order moment M [sigma] anda normalized RCS second-order moment [upsilon] [sigma]; 5, multiplying the weights H1, H2, H3 and H4 by the normalized RCS first order moment, the normalized RCS second-order moment, a spectral peak number and a peak speed respectively, and performing calculation to obtain a probability P < person > of judging the current X times of tracking data as a person and a probability P < vehicle > of judging the current X times of tracking data as a vehicle; 6, judging that the target is a person if the P person value is greater than the P vehicle value; otherwise, judging that the target is a vehicle, and outputting a judgment result. According to the identification method, as long as the radar erection posture is correct and no shielding exists between the target and the radar, high identification probability can be realized, and stable and reliable identification can be realized.