Extreme high temperature (EHT) events have caused serious impact on society. In previous research, EHT indices seldom consider spatial aggregation. To conduct the comprehensive evaluation, the global EHT days from 1979 to 2017 have been evaluated in terms of high temperature duration, accumulative temperature intensities and magnitude of spatial aggregation based on ERA‐Interim 2 m air temperature data set. In addition, abnormal warming of polar and ocean surfaces has rarely been addressed in global‐scale studies. To address this deficiency, the indices of EHT days are normalized by maximum in history for each location. To solve the time discontinuity problem on EHT spatiotemporal distribution in previous studies, we utilized spatiotemporal scan statistics to detect EHT spatiotemporal clusters (EHT‐STCs), and clusters correspond to a certain spatiotemporal extent where the EHT events are aggregated. This method actualizes time‐continuous and spatiotemporal integral detection. The results show the following: 1. There are apparent changes in EHT‐STC spatial distribution after 1999, so 1999 can be recognized as the critical time node of EHT‐STC spatial transformation. 2. After 1999, the EHT‐STC number gradually increased, and the duration and spatial coverage of EHT‐STC gradually expanded. Overall, EHT‐STCs display a significant northward migration, with an average northward movement of 7.03° for each transition from 1979 to 2017. 3. There are differences in the extremity and transformation process in different regions. EHT‐STCs in Antarctica are concentrated before 1992, with the strongest EHT extremity; EHT‐STCs in the Arctic and middle‐low latitude continents are concentrated after 2010 and 1999 respectively, both with a strong EHT extremity. There is a similar EHT‐STC number in the middle‐low latitude ocean before and after 1999 with the weakest extremity. These results contribute to determining EHT causes and future trends in global climate change.