Abstract: In order to discriminate neutron and γ ray with high performance, a digital nuclear pulse sampling system based on a Cs₂LiYCl₆: Ce³⁺ (CLYC) detector and a digital oscilloscope was set up. By Matlab programming, two pulse shape discrimination methods, including hierarchical clustering method and K-means clustering method were studied, and were compared with the traditional charge comparison method. The results show that both of these two methods can discriminate neutron and γ pulses accurately. The K-means clustering method is better for less computing time and memory consumption, and makes profit for real-time pulse discriminating. This research provides a useful pulse shape discrimination solution to develop the n/γ dual-mode spectrometer with CLYC detector.