The invention discloses an unmanned aerial vehicle identification method and system based on real-time signal feature analysis. The method comprises the steps of S10, preprocessing received original signals, thereby obtaining time continuous real-time signals; S20, carrying out signal feature parameter extraction, carrying out classification and identification and determining signal parameters; and S30, carrying out analysis and statistics on sorting results, thereby judging whether unmanned aerial vehicles exist in the current signals or not. The S20 comprises the process of carrying out noise base judgment on the received real-time signals, calculating initiation and termination frequencies, central frequencies and occupied bandwidths of the signals according to a noise threshold; addingtime labels to the signals; and acquiring frequency hopping signal features according to features comprising time, frequencies and bandwidths of the signals at each frequency point. According to themethod and the system, the signal analysis is carried out based on continuous time signals, through deep analysis of the signal features, relatively much information can be mined, and the intercept probability of unmanned aerial vehicle signals can be effectively improved.