With the advancement of satellite remote sensing technology, thermal infrared anomaly monitoring has become a critical technique for identifying earthquake precursors. Outgoing longwave radiation (OLR), surface temperature field, pressure field, and snow cover rate data from six M_S ≥ 6.5 earthquakes on the Qinghai-Xizang Plateau (2010—2023) were integrated, and wavelet transform and power spectrum analysis were applied to reveal pre-earthquake thermodynamic anomaly patterns. The results show that the OLR power spectrum exhibits significant anomalies within three months prior to strong earthquakes. The amplitude and scope of these anomalies are positively correlated with the earthquake magnitude. Spatially, three patterns including epicenter-overlying are observed, with most anomalies occurring within 120 km around the epicenter. Some cases present precursory double-peak anomalies, which provide physical evidence for predicting the rupture direction. The surface air temperature follows a three-stage pattern of “warming-cooling-re-warming”, which is consistent with the spatiotemporal variations of OLR. The snow cover ratio indirectly affects the OLR intensity by regulating the surface albedo and thus can serve as an auxiliary criterion, while the pressure field anomalies show no significant correlation with earthquake precursors. The lithosphere-atmosphere-ionosphere (LAI) coupling theory and other related mechanisms can explain the physical causes of such thermodynamic anomalies, which lay a foundation for establishing discriminant models between earthquake types and thermal anomalies and provide theoretical support for promoting the short-term prediction of strong earthquakes.