Abstract: With the rapid development of drone technology, its application areas are constantly expanding and widely used in military, civil and commercial fields. Drone identification systems play an important role in ensuring the legal flight of drones and resisting potential threats. However, the existing drone detection and identification schemes have shortcomings such as low accuracy, poor robustness, complex data processing and susceptibility to environmental influences when facing small, low-altitude, low-speed or hovering drones. Therefore, a drone detection and identification system based on drone noise fingerprint is proposed. In the data preprocessing stage, the system first removes the interference of environmental noise and performs data enhancement on the training audio data to enrich the diversity of the training data set and adapt to the noise characteristics of drones under different operating conditions. Next, the feature extraction module reduces the instability of features caused by changes in signal strength or distance when the system extracts features by designing a time domain peak normalization algorithm and a feature vector rescaling algorithm, thereby improving the distance robustness of the system. Finally, the system constructs a new drone identification algorithm based on multiple KNN classifiers to realize the identification of drone models. Experimental results show that the system can accurately detect and identify drones under a variety of different environments and distance conditions.