Abstract: (Objective) Nuclear explosion monitoring is a critical technology for nuclear test ban verification. In order to monitor potential nuclear tests worldwide, the Comprehensive Nuclear-Test-Ban Treaty (CTBT) establishes a rigorous verification regime. Within this framework, waveform data from the International Monitoring System (IMS) are transmitted in real time to the International Data Centre (IDC) for processing and analyzing. The IDC delivers automated processing results at three stages: approximately 1 hour, 4 hours, and 6 hours after data acquisition. For region-specific nuclear explosion monitoring, direct reliance on IDC results faces challenges of delayed response and high false detection rates. (Method) To address these limitations, we propose Seisprint, a rapid detection method based on seismic waveform fingerprints. Inspired by audio fingerprinting techniques, Seisprint utilizes historical nuclear explosion waveforms as templates. Continuous seismic waveforms are compressed into multiple binary fingerprints through sliding-window feature extraction. Automated nuclear event detection and real-time alerts are achieved via rapid similarity matching and clustering. The method was tested using data from two IMS seismic stations near North Korea and four seismic stations in northeastern China, covering six historical underground nuclear tests and natural seismic events in North Korea. (Results) The fingerprints generated by Seisprint effectively distinguish nuclear explosions from non-nuclear signals and demonstrate strong robustness against noise. The method processes an entire day of continuous data from multiple seismic stations within just a few minutes, enabling rapid detection of underground nuclear explosion events. (Conclusion) Results confirm that Seisprint promote the timeliness and accuracy of underground nuclear explosion detection in specific area.