The land surface temperature (LST) in permafrost regions in the Northeast permafrost on March 22 (spring), June 24 (summer), September 21 (autumn), and December 24 (winter) in 2019 were retrieved based on the AMSR2 brightness temperature data. An in-depth analysis of the temperature retrieval accuracy between different types of frozen ground, vegetation cover, and during the four seasons of the day or night was conducted. The results show that: (1) The retrieval accuracy of the four seasons lowers in the seasonal order of summer > autumn > spring > winter, and the accuracy of data at the night was better than that of the day; (2) The retrieval accuracy of different vegetation cover types lowers in the order of grassland > agricultural land > forest land, and; (3) The retrieval accuracy of different frozen ground types lowers in the order of the zone of seasonal frost > zone of isolated patches of thawing permafrost > zone of island permafrost zone > continuous permafrost zone.

基于AMSR2亮温数据反演2019年3月22日（春季）、6月24日（夏季）、9月21日（秋季）、12月24日（冬季）东北多年冻土区多年冻土区地表温度（LST） . 对不同类型冻土、植被覆盖以及昼夜四个季节的温度反演精度进行了深入分析。结果表明：（1）4个季节的检索准确率按夏季>秋季>春季>冬季的季节顺序依次降低，夜间数据的准确率优于白天；(2) 不同植被覆盖类型的检索精度依次为草地>农田>林地，且；