A method based on noise analysis techniques that can be applied to the identification of two-phase flow patterns in nuclear reactors is proposed. The identifying criterion, the high-frequency contribution fraction (HFCF), offers new potential to the in-core recognition of two-phase flow patterns. By analyzing 76 sets of signals acquired from a research nuclear reactor where two-phase flow patterns are generated in an in-core air/water loop, the typical signal, autocorrelogram, and spectrum of each flow pattern are demonstrated and evaluated. The identification success rate is 87 or 93%, depending on whether churn flow is counted. A method to improve the identification rate is also presented. In comparison with our previous work, this study demonstrates that the fluctuation characteristics above 10 Hz are induced by two-phase flow itself and are independent of the driving source; thus, it is adequate to apply the HFCF to the identification of two-phase flow patterns. The present study shows that it is possible to identify two-phase flow patterns by HFCF values.